4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
26 * 'Configure' has been changed to figure out the configuration target if
27 none is given on the command line. Consequently, the 'config' script is
28 now only a mere wrapper. All documentation is changed to only mention
31 *Rich Salz and Richard Levitte*
33 * Added a library context that applications as well as other
34 libraries can use to form a separate context within which libcrypto
35 operations are performed.
37 There are two ways this can be used:
39 - Directly, by passing a library context to functions that take
40 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
42 - Indirectly, by creating a new library context and then assigning
43 it as the new default, with `OPENSSL_CTX_set0_default`.
45 All public OpenSSL functions that take an `OPENSSL_CTX` pointer,
46 apart from the functions directly related to `OPENSSL_CTX`, accept
47 NULL to indicate that the default library context should be used.
49 Library code that changes the default library context using
50 `OPENSSL_CTX_set0_default` should take care to restore it with a
51 second call before returning to the caller.
55 * Handshake now fails if Extended Master Secret extension is dropped
60 * Dropped interactive mode from the 'openssl' program. From now on,
61 the `openssl` command without arguments is equivalent to `openssl
66 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
67 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
68 While the old function names have been retained for backward compatibility
69 they should not be used in new developments
70 because their return values are confusing: Unlike other `_cmp()` functions
71 they do not return 0 in case their arguments are equal.
75 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
76 EC_GROUP_get_field_type().
80 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
81 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
82 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
83 Applications should rely on the library automatically assigning a suitable
84 EC_METHOD internally upon EC_GROUP construction.
88 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
89 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
90 assigned internally without application intervention.
91 Users of EC_GROUP_new() should switch to a different suitable constructor.
95 * Add CAdES-BES signature verification support, mostly derived
96 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
98 *Filipe Raimundo da Silva*
100 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
104 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
105 functions are not widely used and now OpenSSL automatically perform this
106 conversion when needed.
110 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
111 EC_KEY_precompute_mult(). These functions are not widely used and
112 applications should instead switch to named curves which OpenSSL has
113 hardcoded lookup tables for.
117 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
118 should instead use the L<EC_POINT_mul(3)> function.
122 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
123 that are not applicable to the new provider model. Applications should
124 instead use EVP_default_properties_is_fips_enabled() and
125 EVP_default_properties_enable_fips().
129 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
130 is set, an unexpected EOF is ignored, it pretends a close notify was received
131 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
135 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
136 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
137 used and applications should instead use the
138 L<EC_POINT_set_affine_coordinates(3)> and
139 L<EC_POINT_get_affine_coordinates(3)> functions.
143 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
144 arrays to be more easily constructed via a series of utility functions.
145 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
146 the various push functions and finally convert to a passable OSSL_PARAM
147 array using OSSL_PARAM_BLD_to_param().
151 * The security strength of SHA1 and MD5 based signatures in TLS has been
152 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
153 working at the default security level of 1 and instead requires security
154 level 0. The security level can be changed either using the cipher string
155 with @SECLEVEL, or calling SSL_CTX_set_security_level().
159 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
160 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
161 internal keys, if they correspond to one of those built in types.
165 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
166 contain a provider side internal key.
170 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
171 They are old functions that we don't use, and that you could disable with
172 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
176 * Project text documents not yet having a proper file name extension
177 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
178 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
182 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
183 have been converted to Markdown with the goal to produce documents
184 which not only look pretty when viewed online in the browser, but
185 remain well readable inside a plain text editor.
187 To achieve this goal, a 'minimalistic' Markdown style has been applied
188 which avoids formatting elements that interfere too much with the
189 reading flow in the text file. For example, it
191 * avoids [ATX headings][] and uses [setext headings][] instead
192 (which works for `<h1>` and `<h2>` headings only).
193 * avoids [inline links][] and uses [reference links][] instead.
194 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
196 [ATX headings]: https://github.github.com/gfm/#atx-headings
197 [setext headings]: https://github.github.com/gfm/#setext-headings
198 [inline links]: https://github.github.com/gfm/#inline-link
199 [reference links]: https://github.github.com/gfm/#reference-link
200 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
201 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
203 *Matthias St. Pierre*
205 * The test suite is changed to preserve results of each test recipe.
206 A new directory test-runs/ with subdirectories named like the
207 test recipes are created in the build tree for this purpose.
211 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
212 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
213 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
215 *David von Oheimb, Martin Peylo*
217 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
218 The legacy OCSP-focused and only partly documented API is retained.
219 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
223 * BIO_do_connect and BIO_do_handshake have been extended:
224 If domain name resolution yields multiple IP addresses all of them are tried
225 after connect() failures.
229 * All of the low level RSA functions have been deprecated including:
231 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
232 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
233 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
234 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
235 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
236 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
237 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
238 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
239 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
240 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
241 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
242 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
243 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
244 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
245 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
246 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
247 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
248 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
249 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
250 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
251 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
252 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
253 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
254 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
255 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
256 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
257 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
258 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
260 Use of these low level functions has been informally discouraged for a long
261 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
262 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
263 L<EVP_PKEY_decrypt(3)>.
267 * X509 certificates signed using SHA1 are no longer allowed at security
269 In TLS/SSL the default security level is 1. It can be set either
270 using the cipher string with @SECLEVEL, or calling
271 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
272 a call to SSL_CTX_use_certificate() will fail if the security level is not
274 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
275 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
280 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
281 modified to use PKEY APIs. These commands are now in maintenance mode
282 and no new features will be added to them.
286 * The command line utility rsautl has been deprecated.
287 Instead use the pkeyutl program.
291 * The command line utilities genrsa and rsa have been modified to use PKEY
292 APIs These commands are now in maintenance mode and no new features will
297 * All of the low level DH functions have been deprecated including:
299 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
300 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
301 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
302 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
303 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
304 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
305 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
306 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
307 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
308 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
309 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
310 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
311 DH_meth_get_generate_params and DH_meth_set_generate_params.
313 Use of these low level functions has been informally discouraged for a long
314 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
315 and L<EVP_PKEY_derive(3)>.
319 * All of the low level DSA functions have been deprecated including:
321 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
322 DSA_get_default_method, DSA_set_method, DSA_get_method,
323 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
324 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
325 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
326 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
327 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
328 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
329 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
330 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
331 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
332 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
333 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
336 Use of these low level functions has been informally discouraged for a long
337 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
338 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
342 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
343 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
344 This means that applications don't have to look at the curve NID and
345 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
346 However, they still can, that EVP_PKEY_set_alias_type() call acts as
347 a no-op when the EVP_PKEY is already of the given type.
349 Parameter and key generation is also reworked to make it possible
350 to generate EVP_PKEY_SM2 parameters and keys without having to go
351 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
352 However, code that does the latter will still work as before.
356 * Deprecated low level ECDH and ECDSA functions. These include:
358 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
359 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
362 Use of these low level functions has been informally discouraged for a long
363 time. Instead applications should use the EVP_PKEY_derive(3),
364 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
368 * Deprecated the EC_KEY_METHOD functions. These include:
370 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
371 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
372 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
373 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
374 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
375 EC_KEY_METHOD_get_verify.
377 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
381 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
382 and EVP_PKEY_decrypt() instead.
383 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
384 and EVP_PKEY_encrypt() instead.
388 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
389 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
390 a new formulation to include all the things it can be used for,
391 as well as words of caution.
395 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
396 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
400 * All of the low level HMAC functions have been deprecated including:
402 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
403 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
406 Use of these low level functions has been informally discouraged for a long
407 time. Instead applications should use L<EVP_MAC_new_ctx(3)>,
408 L<EVP_MAC_free_ctx(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
409 and L<EVP_MAC_final(3)>.
413 * Over two thousand fixes were made to the documentation, including:
414 - Common options (such as -rand/-writerand, TLS version control, etc)
415 were refactored and point to newly-enhanced descriptions in openssl.pod.
416 - Added style conformance for all options (with help from Richard Levitte),
417 documented all reported missing options, added a CI build to check
418 that all options are documented and that no unimplemented options
420 - Documented some internals, such as all use of environment variables.
421 - Addressed all internal broken L<> references.
425 * All of the low level CMAC functions have been deprecated including:
427 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
428 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
430 Use of these low level functions has been informally discouraged for a long
431 time. Instead applications should use L<EVP_MAC_new_ctx(3)>,
432 L<EVP_MAC_free_ctx(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
433 and L<EVP_MAC_final(3)>.
437 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
438 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
441 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
442 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
443 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
444 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
445 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
446 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
447 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
448 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
449 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
450 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
452 Use of these low level functions has been informally discouraged
453 for a long time. Applications should use the EVP_DigestInit_ex(3),
454 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
458 * Corrected the documentation of the return values from the `EVP_DigestSign*`
459 set of functions. The documentation mentioned negative values for some
460 errors, but this was never the case, so the mention of negative values
463 Code that followed the documentation and thereby check with something
464 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
468 * All of the low level cipher functions have been deprecated including:
470 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
471 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
472 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
473 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
474 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
475 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
476 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
477 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
478 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
479 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
480 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
481 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
482 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
483 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
484 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
485 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
486 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
487 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
488 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
489 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
490 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
491 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
492 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
493 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
494 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
495 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
496 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
497 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
498 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
500 Use of these low level functions has been informally discouraged for
501 a long time. Applications should use the high level EVP APIs, e.g.
502 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
503 equivalently named decrypt functions instead.
505 *Matt Caswell and Paul Dale*
507 * Removed include/openssl/opensslconf.h.in and replaced it with
508 include/openssl/configuration.h.in, which differs in not including
509 <openssl/macros.h>. A short header include/openssl/opensslconf.h
510 was added to include both.
512 This allows internal hacks where one might need to modify the set
513 of configured macros, for example this if deprecated symbols are
514 still supposed to be available internally:
516 #include <openssl/configuration.h>
518 #undef OPENSSL_NO_DEPRECATED
519 #define OPENSSL_SUPPRESS_DEPRECATED
521 #include <openssl/macros.h>
523 This should not be used by applications that use the exported
524 symbols, as that will lead to linking errors.
528 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
529 used in exponentiation with 512-bit moduli. No EC algorithms are
530 affected. Analysis suggests that attacks against 2-prime RSA1024,
531 3-prime RSA1536, and DSA1024 as a result of this defect would be very
532 difficult to perform and are not believed likely. Attacks against DH512
533 are considered just feasible. However, for an attack the target would
534 have to re-use the DH512 private key, which is not recommended anyway.
535 Also applications directly using the low level API BN_mod_exp may be
536 affected if they use BN_FLG_CONSTTIME.
541 * Most memory-debug features have been deprecated, and the functionality
542 replaced with no-ops.
546 * Added documentation for the STACK API. OpenSSL only defines the STACK
547 functions where they are used.
551 * Introduced a new method type and API, OSSL_SERIALIZER, to
552 represent generic serializers. An implementation is expected to
553 be able to serialize an object associated with a given name (such
554 as an algorithm name for an asymmetric key) into forms given by
555 implementation properties.
557 Serializers are primarily used from inside libcrypto, through
558 calls to functions like EVP_PKEY_print_private(),
559 PEM_write_bio_PrivateKey() and similar.
561 Serializers are specified in such a way that they can be made to
562 directly handle the provider side portion of an object, if this
563 provider side part comes from the same provider as the serializer
564 itself, but can also be made to handle objects in parametrized
565 form (as an OSSL_PARAM array of data). This allows a provider to
566 offer generic serializers as a service for any other provider.
570 * Added a .pragma directive to the syntax of configuration files, to
571 allow varying behavior in a supported and predictable manner.
572 Currently added pragma:
576 This allows dollar signs to be a keyword character unless it's
577 followed by a opening brace or parenthesis. This is useful for
578 platforms where dollar signs are commonly used in names, such as
579 volume names and system directory names on VMS.
583 * Added functionality to create an EVP_PKEY from user data. This
584 is effectively the same as creating a RSA, DH or DSA object and
585 then assigning them to an EVP_PKEY, but directly using algorithm
586 agnostic EVP functions. A benefit is that this should be future
587 proof for public key algorithms to come.
591 * Change the interpretation of the '--api' configuration option to
592 mean that this is a desired API compatibility level with no
593 further meaning. The previous interpretation, that this would
594 also mean to remove all deprecated symbols up to and including
595 the given version, no requires that 'no-deprecated' is also used
596 in the configuration.
598 When building applications, the desired API compatibility level
599 can be set with the OPENSSL_API_COMPAT macro like before. For
600 API compatibility version below 3.0, the old style numerical
601 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
602 For version 3.0 and on, the value is expected to be the decimal
603 value calculated from the major and minor version like this:
605 MAJOR * 10000 + MINOR * 100
609 -DOPENSSL_API_COMPAT=30000 For 3.0
610 -DOPENSSL_API_COMPAT=30200 For 3.2
612 To hide declarations that are deprecated up to and including the
613 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
614 given when building the application as well.
618 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
619 access to certificate and CRL stores via URIs and OSSL_STORE
622 This adds the following functions:
624 - X509_LOOKUP_store()
625 - X509_STORE_load_file()
626 - X509_STORE_load_path()
627 - X509_STORE_load_store()
628 - SSL_add_store_cert_subjects_to_stack()
629 - SSL_CTX_set_default_verify_store()
630 - SSL_CTX_load_verify_file()
631 - SSL_CTX_load_verify_dir()
632 - SSL_CTX_load_verify_store()
636 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
637 The presence of this system service is determined at run-time.
641 * Added functionality to create an EVP_PKEY context based on data
642 for methods from providers. This takes an algorithm name and a
643 property query string and simply stores them, with the intent
644 that any operation that uses this context will use those strings
645 to fetch the needed methods implicitly, thereby making the port
646 of application written for pre-3.0 OpenSSL easier.
650 * The undocumented function NCONF_WIN32() has been deprecated; for
651 conversion details see the HISTORY section of doc/man5/config.pod
655 * Introduced the new functions EVP_DigestSignInit_ex() and
656 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
657 EVP_DigestVerifyUpdate() have been converted to functions. See the man
658 pages for further details.
662 * Over two thousand fixes were made to the documentation, including:
663 adding missing command flags, better style conformance, documentation
666 *Rich Salz, Richard Levitte*
668 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
669 X25519, X448, Ed25519 and Ed448.
673 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
678 * Deprecated the public definition of ERR_STATE as well as the function
679 ERR_get_state(). This is done in preparation of making ERR_STATE an
684 * Added ERR functionality to give callers access to the stored function
685 names that have replaced the older function code based functions.
687 New functions are ERR_get_error_func(), ERR_peek_error_func(),
688 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
689 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
690 and ERR_peek_last_error_all().
692 These functions have become deprecated: ERR_get_error_line_data(),
693 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
694 ERR_func_error_string().
698 * Extended testing to be verbose for failing tests only. The make variables
699 VERBOSE_FAILURE or VF can be used to enable this:
701 $ make VF=1 test # Unix
702 $ mms /macro=(VF=1) test ! OpenVMS
703 $ nmake VF=1 test # Windows
707 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
708 used even when parsing explicit parameters, when loading a serialized key
709 or calling `EC_GROUP_new_from_ecpkparameters()`/
710 `EC_GROUP_new_from_ecparameters()`.
711 This prevents bypass of security hardening and performance gains,
712 especially for curves with specialized EC_METHODs.
713 By default, if a key encoded with explicit parameters is loaded and later
714 serialized, the output is still encoded with explicit parameters, even if
715 internally a "named" EC_GROUP is used for computation.
719 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
720 this change, EC_GROUP_set_generator would accept order and/or cofactor as
721 NULL. After this change, only the cofactor parameter can be NULL. It also
722 does some minimal sanity checks on the passed order.
727 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
728 An attack is simple, if the first CMS_recipientInfo is valid but the
729 second CMS_recipientInfo is chosen ciphertext. If the second
730 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
731 encryption key will be replaced by garbage, and the message cannot be
732 decoded, but if the RSA decryption fails, the correct encryption key is
733 used and the recipient will not notice the attack.
734 As a work around for this potential attack the length of the decrypted
735 key must be equal to the cipher default key length, in case the
736 certifiate is not given and all recipientInfo are tried out.
737 The old behaviour can be re-enabled in the CMS code by setting the
738 CMS_DEBUG_DECRYPT flag.
742 * Early start up entropy quality from the DEVRANDOM seed source has been
743 improved for older Linux systems. The RAND subsystem will wait for
744 /dev/random to be producing output before seeding from /dev/urandom.
745 The seeded state is stored for future library initialisations using
746 a system global shared memory segment. The shared memory identifier
747 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
748 the desired value. The default identifier is 114.
752 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
753 when primes for RSA keys are computed.
754 Since we previously always generated primes == 2 (mod 3) for RSA keys,
755 the 2-prime and 3-prime RSA modules were easy to distinguish, since
756 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
757 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
758 This avoids possible fingerprinting of newly generated RSA modules.
762 * Correct the extended master secret constant on EBCDIC systems. Without this
763 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
764 negotiate EMS will fail. Unfortunately this also means that TLS connections
765 between EBCDIC systems with this fix, and EBCDIC systems without this
766 fix will fail if they negotiate EMS.
770 * Changed the library initialisation so that the config file is now loaded
771 by default. This was already the case for libssl. It now occurs for both
772 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
773 OPENSSL_init_crypto() to suppress automatic loading of a config file.
777 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
778 where the former acts as a replacement for ERR_put_error(), and the
779 latter replaces the combination ERR_put_error()+ERR_add_error_data().
780 ERR_raise_data() adds more flexibility by taking a format string and
781 an arbitrary number of arguments following it, to be processed with
786 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
787 to check if a named provider is loaded and available. When called, it
788 will also activate all fallback providers if such are still present.
792 * Enforce a minimum DH modulus size of 512 bits.
796 * Changed DH parameters to generate the order q subgroup instead of 2q.
797 Previously generated DH parameters are still accepted by DH_check
798 but DH_generate_key works around that by clearing bit 0 of the
799 private key for those. This avoids leaking bit 0 of the private key.
803 * Significantly reduce secure memory usage by the randomness pools.
807 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
812 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
813 algorithms. An implementation of a key exchange algorithm can be obtained
814 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
815 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
816 the older EVP_PKEY_derive_init() function. See the man pages for the new
817 functions for further details.
821 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
825 * Removed the function names from error messages and deprecated the
828 * Removed NextStep support and the macro OPENSSL_UNISTD
832 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
833 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
834 Also removed "export var as function" capability; we do not export
835 variables, only functions.
839 * RC5_32_set_key has been changed to return an int type, with 0 indicating
840 an error and 1 indicating success. In previous versions of OpenSSL this
841 was a void type. If a key was set longer than the maximum possible this
846 * Support SM2 signing and verification schemes with X509 certificate.
850 * Use SHA256 as the default digest for TS query in the ts app.
854 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
855 This checks that the salt length is at least 128 bits, the derived key
856 length is at least 112 bits, and that the iteration count is at least 1000.
857 For backwards compatibility these checks are disabled by default in the
858 default provider, but are enabled by default in the fips provider.
859 To enable or disable these checks use the control
860 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
864 * Default cipher lists/suites are now available via a function, the
865 #defines are deprecated.
869 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
870 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
871 for Windows Store apps easier. Also, the "no-uplink" option has been added.
875 * Join the directories crypto/x509 and crypto/x509v3
879 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
880 This changes the size when using the genpkey app when no size is given. It
881 fixes an omission in earlier changes that changed all RSA, DSA and DH
882 generation apps to use 2048 bits by default.
886 * Added command 'openssl kdf' that uses the EVP_KDF API.
890 * Added command 'openssl mac' that uses the EVP_MAC API.
894 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
895 as default directories. Also added the command 'openssl info'
896 for scripting purposes.
900 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
901 deprecated. These undocumented functions were never integrated into the EVP
902 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
903 Bi-directional IGE mode. These modes were never formally standardised and
904 usage of these functions is believed to be very small. In particular
905 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
906 is ever used. The security implications are believed to be minimal, but
907 this issue was never fixed for backwards compatibility reasons. New code
908 should not use these modes.
912 * Add prediction resistance to the DRBG reseeding process.
916 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
917 mandated by IEEE Std 1619-2018.
921 * Added newline escaping functionality to a filename when using openssl dgst.
922 This output format is to replicate the output format found in the `*sum`
923 checksum programs. This aims to preserve backward compatibility.
925 *Matt Eaton, Richard Levitte, and Paul Dale*
927 * Removed the heartbeat message in DTLS feature, as it has very
928 little usage and doesn't seem to fulfill a valuable purpose.
929 The configuration option is now deprecated.
933 * Changed the output of 'openssl {digestname} < file' to display the
934 digest name in its output.
938 * Added a new generic trace API which provides support for enabling
939 instrumentation through trace output. This feature is mainly intended
940 as an aid for developers and is disabled by default. To utilize it,
941 OpenSSL needs to be configured with the `enable-trace` option.
943 If the tracing API is enabled, the application can activate trace output
944 by registering BIOs as trace channels for a number of tracing and debugging
947 The 'openssl' application has been expanded to enable any of the types
948 available via environment variables defined by the user, and serves as
949 one possible example on how to use this functionality.
951 *Richard Levitte & Matthias St. Pierre*
953 * Added build tests for C++. These are generated files that only do one
954 thing, to include one public OpenSSL head file each. This tests that
955 the public header files can be usefully included in a C++ application.
957 This test isn't enabled by default. It can be enabled with the option
958 'enable-buildtest-c++'.
962 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
966 * Add KMAC to EVP_MAC.
970 * Added property based algorithm implementation selection framework to
975 * Added SCA hardening for modular field inversion in EC_GROUP through
976 a new dedicated field_inv() pointer in EC_METHOD.
977 This also addresses a leakage affecting conversions from projective
978 to affine coordinates.
980 *Billy Bob Brumley, Nicola Tuveri*
982 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
983 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
984 those algorithms that were already supported through the EVP_PKEY API
985 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
986 and scrypt are now wrappers that call EVP_KDF.
990 * Build devcrypto engine as a dynamic engine.
994 * Add keyed BLAKE2 to EVP_MAC.
998 * Fix a bug in the computation of the endpoint-pair shared secret used
999 by DTLS over SCTP. This breaks interoperability with older versions
1000 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1001 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1002 interoperability with such broken implementations. However, enabling
1003 this switch breaks interoperability with correct implementations.
1005 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1006 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1010 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1014 * Change the license to the Apache License v2.0.
1018 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1020 - Major releases (indicated by incrementing the MAJOR release number)
1021 may introduce incompatible API/ABI changes.
1022 - Minor releases (indicated by incrementing the MINOR release number)
1023 may introduce new features but retain API/ABI compatibility.
1024 - Patch releases (indicated by incrementing the PATCH number)
1025 are intended for bug fixes and other improvements of existing
1026 features only (like improving performance or adding documentation)
1027 and retain API/ABI compatibility.
1031 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1035 * Remove the 'dist' target and add a tarball building script. The
1036 'dist' target has fallen out of use, and it shouldn't be
1037 necessary to configure just to create a source distribution.
1041 * Recreate the OS390-Unix config target. It no longer relies on a
1042 special script like it did for OpenSSL pre-1.1.0.
1046 * Instead of having the source directories listed in Configure, add
1047 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1052 * Add GMAC to EVP_MAC.
1056 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1060 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1061 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1062 to facilitate the continued use of MACs through raw private keys in
1063 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1067 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1068 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1072 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1073 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1074 are retained for backwards compatibility.
1078 * AES-XTS mode now enforces that its two keys are different to mitigate
1079 the attacked described in "Efficient Instantiations of Tweakable
1080 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1081 Details of this attack can be obtained from:
1082 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1086 * Rename the object files, i.e. give them other names than in previous
1087 versions. Their names now include the name of the final product, as
1088 well as its type mnemonic (bin, lib, shlib).
1092 * Added new option for 'openssl list', '-objects', which will display the
1093 list of built in objects, i.e. OIDs with names.
1097 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1098 improves application performance by removing data copies and providing
1099 applications with zero-copy system calls such as sendfile and splice.
1103 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1104 option is set, openssl cleanses (zeroize) plaintext bytes from
1105 internal buffers after delivering them to the application. Note,
1106 the application is still responsible for cleansing other copies
1107 (e.g.: data received by SSL_read(3)).
1114 ### Changes between 1.1.1e and 1.1.1f [xx XXX xxxx]
1116 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1118 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1119 while reading in libssl then we would report an error back to the
1120 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1121 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1122 therefore give a hint as to what went wrong.
1126 * Check that ed25519 and ed448 are allowed by the security level. Previously
1127 signature algorithms not using an MD were not being checked that they were
1128 allowed by the security level.
1132 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1133 was not quite right. The behaviour was not consistent between resumption
1134 and normal handshakes, and also not quite consistent with historical
1135 behaviour. The behaviour in various scenarios has been clarified and
1136 it has been updated to make it match historical behaviour as closely as
1141 * *[VMS only]* The header files that the VMS compilers include automatically,
1142 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1143 that the C++ compiler doesn't understand. This is a shortcoming in the
1144 compiler, but can be worked around with `__cplusplus` guards.
1146 C++ applications that use OpenSSL libraries must be compiled using the
1147 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1148 functions. Otherwise, only functions with symbols of less than 31
1149 characters can be used, as the linker will not be able to successfully
1150 resolve symbols with longer names.
1154 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1155 The presence of this system service is determined at run-time.
1159 * Added newline escaping functionality to a filename when using openssl dgst.
1160 This output format is to replicate the output format found in the `*sum`
1161 checksum programs. This aims to preserve backward compatibility.
1163 *Matt Eaton, Richard Levitte, and Paul Dale*
1165 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1170 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1172 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1173 number generator (RNG). This was intended to include protection in the
1174 event of a fork() system call in order to ensure that the parent and child
1175 processes did not share the same RNG state. However this protection was not
1176 being used in the default case.
1178 A partial mitigation for this issue is that the output from a high
1179 precision timer is mixed into the RNG state so the likelihood of a parent
1180 and child process sharing state is significantly reduced.
1182 If an application already calls OPENSSL_init_crypto() explicitly using
1183 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1186 *Matthias St. Pierre*
1188 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1189 used even when parsing explicit parameters, when loading a serialized key
1190 or calling `EC_GROUP_new_from_ecpkparameters()`/
1191 `EC_GROUP_new_from_ecparameters()`.
1192 This prevents bypass of security hardening and performance gains,
1193 especially for curves with specialized EC_METHODs.
1194 By default, if a key encoded with explicit parameters is loaded and later
1195 serialized, the output is still encoded with explicit parameters, even if
1196 internally a "named" EC_GROUP is used for computation.
1200 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1201 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1202 NULL. After this change, only the cofactor parameter can be NULL. It also
1203 does some minimal sanity checks on the passed order.
1208 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1209 An attack is simple, if the first CMS_recipientInfo is valid but the
1210 second CMS_recipientInfo is chosen ciphertext. If the second
1211 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1212 encryption key will be replaced by garbage, and the message cannot be
1213 decoded, but if the RSA decryption fails, the correct encryption key is
1214 used and the recipient will not notice the attack.
1215 As a work around for this potential attack the length of the decrypted
1216 key must be equal to the cipher default key length, in case the
1217 certifiate is not given and all recipientInfo are tried out.
1218 The old behaviour can be re-enabled in the CMS code by setting the
1219 CMS_DEBUG_DECRYPT flag.
1224 * Early start up entropy quality from the DEVRANDOM seed source has been
1225 improved for older Linux systems. The RAND subsystem will wait for
1226 /dev/random to be producing output before seeding from /dev/urandom.
1227 The seeded state is stored for future library initialisations using
1228 a system global shared memory segment. The shared memory identifier
1229 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1230 the desired value. The default identifier is 114.
1234 * Correct the extended master secret constant on EBCDIC systems. Without this
1235 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1236 negotiate EMS will fail. Unfortunately this also means that TLS connections
1237 between EBCDIC systems with this fix, and EBCDIC systems without this
1238 fix will fail if they negotiate EMS.
1242 * Use Windows installation paths in the mingw builds
1244 Mingw isn't a POSIX environment per se, which means that Windows
1245 paths should be used for installation.
1250 * Changed DH_check to accept parameters with order q and 2q subgroups.
1251 With order 2q subgroups the bit 0 of the private key is not secret
1252 but DH_generate_key works around that by clearing bit 0 of the
1253 private key for those. This avoids leaking bit 0 of the private key.
1257 * Significantly reduce secure memory usage by the randomness pools.
1261 * Revert the DEVRANDOM_WAIT feature for Linux systems
1263 The DEVRANDOM_WAIT feature added a select() call to wait for the
1264 /dev/random device to become readable before reading from the
1265 /dev/urandom device.
1267 It turned out that this change had negative side effects on
1268 performance which were not acceptable. After some discussion it
1269 was decided to revert this feature and leave it up to the OS
1270 resp. the platform maintainer to ensure a proper initialization
1271 during early boot time.
1273 *Matthias St. Pierre*
1275 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1277 * Add build tests for C++. These are generated files that only do one
1278 thing, to include one public OpenSSL head file each. This tests that
1279 the public header files can be usefully included in a C++ application.
1281 This test isn't enabled by default. It can be enabled with the option
1282 'enable-buildtest-c++'.
1286 * Enable SHA3 pre-hashing for ECDSA and DSA.
1290 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1291 This changes the size when using the genpkey app when no size is given. It
1292 fixes an omission in earlier changes that changed all RSA, DSA and DH
1293 generation apps to use 2048 bits by default.
1297 * Reorganize the manual pages to consistently have RETURN VALUES,
1298 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1299 util/fix-doc-nits accordingly.
1301 *Paul Yang, Joshua Lock*
1303 * Add the missing accessor EVP_PKEY_get0_engine()
1307 * Have apps like 's_client' and 's_server' output the signature scheme
1308 along with other cipher suite parameters when debugging.
1312 * Make OPENSSL_config() error agnostic again.
1316 * Do the error handling in RSA decryption constant time.
1320 * Prevent over long nonces in ChaCha20-Poly1305.
1322 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1323 for every encryption operation. RFC 7539 specifies that the nonce value
1324 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1325 and front pads the nonce with 0 bytes if it is less than 12
1326 bytes. However it also incorrectly allows a nonce to be set of up to 16
1327 bytes. In this case only the last 12 bytes are significant and any
1328 additional leading bytes are ignored.
1330 It is a requirement of using this cipher that nonce values are
1331 unique. Messages encrypted using a reused nonce value are susceptible to
1332 serious confidentiality and integrity attacks. If an application changes
1333 the default nonce length to be longer than 12 bytes and then makes a
1334 change to the leading bytes of the nonce expecting the new value to be a
1335 new unique nonce then such an application could inadvertently encrypt
1336 messages with a reused nonce.
1338 Additionally the ignored bytes in a long nonce are not covered by the
1339 integrity guarantee of this cipher. Any application that relies on the
1340 integrity of these ignored leading bytes of a long nonce may be further
1341 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1342 is safe because no such use sets such a long nonce value. However user
1343 applications that use this cipher directly and set a non-default nonce
1344 length to be longer than 12 bytes may be vulnerable.
1346 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1352 * Add DEVRANDOM_WAIT feature for Linux systems
1354 On older Linux systems where the getrandom() system call is not available,
1355 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1356 Contrary to getrandom(), the /dev/urandom device will not block during
1357 early boot when the kernel CSPRNG has not been seeded yet.
1359 To mitigate this known weakness, use select() to wait for /dev/random to
1360 become readable before reading from /dev/urandom.
1362 * Ensure that SM2 only uses SM3 as digest algorithm
1366 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1368 * Change the info callback signals for the start and end of a post-handshake
1369 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1370 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1371 confused by this and assume that a TLSv1.2 renegotiation has started. This
1372 can break KeyUpdate handling. Instead we no longer signal the start and end
1373 of a post handshake message exchange (although the messages themselves are
1374 still signalled). This could break some applications that were expecting
1375 the old signals. However without this KeyUpdate is not usable for many
1380 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1382 * Timing vulnerability in DSA signature generation
1384 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1385 timing side channel attack. An attacker could use variations in the signing
1386 algorithm to recover the private key.
1388 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1393 * Timing vulnerability in ECDSA signature generation
1395 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1396 timing side channel attack. An attacker could use variations in the signing
1397 algorithm to recover the private key.
1399 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1404 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1405 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1406 of two gigabytes and the error handling improved.
1408 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1409 categorized as a normal bug, not a security issue, because the DRBG reseeds
1410 automatically and is fully functional even without additional randomness
1411 provided by the application.
1413 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1415 * Add a new ClientHello callback. Provides a callback interface that gives
1416 the application the ability to adjust the nascent SSL object at the
1417 earliest stage of ClientHello processing, immediately after extensions have
1418 been collected but before they have been processed. In particular, this
1419 callback can adjust the supported TLS versions in response to the contents
1424 * Add SM2 base algorithm support.
1428 * s390x assembly pack: add (improved) hardware-support for the following
1429 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1430 aes-cfb/cfb8, aes-ecb.
1434 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1435 parameter is no longer accepted, as it leads to a corrupt table. NULL
1436 pem_str is reserved for alias entries only.
1440 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1441 step for prime curves. The new implementation is based on formulae from
1442 differential addition-and-doubling in homogeneous projective coordinates
1443 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1444 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1445 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1446 to work in projective coordinates.
1448 *Billy Bob Brumley, Nicola Tuveri*
1450 * Change generating and checking of primes so that the error rate of not
1451 being prime depends on the intended use based on the size of the input.
1452 For larger primes this will result in more rounds of Miller-Rabin.
1453 The maximal error rate for primes with more than 1080 bits is lowered
1456 *Kurt Roeckx, Annie Yousar*
1458 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1462 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1463 moving between systems, and to avoid confusion when a Windows build is
1464 done with mingw vs with MSVC. For POSIX installs, there's still a
1465 symlink or copy named 'tsget' to avoid that confusion as well.
1469 * Revert blinding in ECDSA sign and instead make problematic addition
1470 length-invariant. Switch even to fixed-length Montgomery multiplication.
1474 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1475 step for binary curves. The new implementation is based on formulae from
1476 differential addition-and-doubling in mixed Lopez-Dahab projective
1477 coordinates, modified to independently blind the operands.
1479 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1481 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1482 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1483 EC_METHODs to implement their own specialized "ladder step", to take
1484 advantage of more favorable coordinate systems or more efficient
1485 differential addition-and-doubling algorithms.
1487 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1489 * Modified the random device based seed sources to keep the relevant
1490 file descriptors open rather than reopening them on each access.
1491 This allows such sources to operate in a chroot() jail without
1492 the associated device nodes being available. This behaviour can be
1493 controlled using RAND_keep_random_devices_open().
1497 * Numerous side-channel attack mitigations have been applied. This may have
1498 performance impacts for some algorithms for the benefit of improved
1499 security. Specific changes are noted in this change log by their respective
1504 * AIX shared library support overhaul. Switch to AIX "natural" way of
1505 handling shared libraries, which means collecting shared objects of
1506 different versions and bitnesses in one common archive. This allows to
1507 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1508 doesn't affect the way 3rd party applications are linked, only how
1509 multi-version installation is managed.
1513 * Make ec_group_do_inverse_ord() more robust and available to other
1514 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1515 mitigations are applied to the fallback BN_mod_inverse().
1516 When using this function rather than BN_mod_inverse() directly, new
1517 EC cryptosystem implementations are then safer-by-default.
1521 * Add coordinate blinding for EC_POINT and implement projective
1522 coordinate blinding for generic prime curves as a countermeasure to
1523 chosen point SCA attacks.
1525 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1527 * Add blinding to ECDSA and DSA signatures to protect against side channel
1528 attacks discovered by Keegan Ryan (NCC Group).
1532 * Enforce checking in the pkeyutl command line app to ensure that the input
1533 length does not exceed the maximum supported digest length when performing
1534 a sign, verify or verifyrecover operation.
1538 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1539 I/O in combination with something like select() or poll() will hang. This
1540 can be turned off again using SSL_CTX_clear_mode().
1541 Many applications do not properly handle non-application data records, and
1542 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1543 around the problems in those applications, but can also break some.
1544 It's recommended to read the manpages about SSL_read(), SSL_write(),
1545 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1546 SSL_CTX_set_read_ahead() again.
1550 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1551 now allow empty (zero character) pass phrases.
1555 * Apply blinding to binary field modular inversion and remove patent
1556 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1560 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1561 binary and prime elliptic curves.
1565 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1566 constant time fixed point multiplication.
1570 * Revise elliptic curve scalar multiplication with timing attack
1571 defenses: ec_wNAF_mul redirects to a constant time implementation
1572 when computing fixed point and variable point multiplication (which
1573 in OpenSSL are mostly used with secret scalars in keygen, sign,
1574 ECDH derive operations).
1575 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1578 * Updated CONTRIBUTING
1582 * Updated DRBG / RAND to request nonce and additional low entropy
1583 randomness from the system.
1585 *Matthias St. Pierre*
1587 * Updated 'openssl rehash' to use OpenSSL consistent default.
1591 * Moved the load of the ssl_conf module to libcrypto, which helps
1592 loading engines that libssl uses before libssl is initialised.
1596 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1600 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1602 *Ingo Schwarze, Rich Salz*
1604 * Added output of accepting IP address and port for 'openssl s_server'
1608 * Added a new API for TLSv1.3 ciphersuites:
1609 SSL_CTX_set_ciphersuites()
1610 SSL_set_ciphersuites()
1614 * Memory allocation failures consistently add an error to the error
1619 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1620 in libcrypto when run as setuid/setgid.
1624 * Load any config file by default when libssl is used.
1628 * Added new public header file <openssl/rand_drbg.h> and documentation
1629 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1631 *Matthias St. Pierre*
1633 * QNX support removed (cannot find contributors to get their approval
1634 for the license change).
1638 * TLSv1.3 replay protection for early data has been implemented. See the
1639 SSL_read_early_data() man page for further details.
1643 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1644 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1645 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1646 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1647 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1648 configuration has been separated out. See the ciphers man page or the
1649 SSL_CTX_set_ciphersuites() man page for more information.
1653 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1654 in responder mode now supports the new "-multi" option, which
1655 spawns the specified number of child processes to handle OCSP
1656 requests. The "-timeout" option now also limits the OCSP
1657 responder's patience to wait to receive the full client request
1658 on a newly accepted connection. Child processes are respawned
1659 as needed, and the CA index file is automatically reloaded
1660 when changed. This makes it possible to run the "ocsp" responder
1661 as a long-running service, making the OpenSSL CA somewhat more
1662 feature-complete. In this mode, most diagnostic messages logged
1663 after entering the event loop are logged via syslog(3) rather than
1668 * Added support for X448 and Ed448. Heavily based on original work by
1673 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1674 objects loaded. This adds the functions OSSL_STORE_expect() and
1675 OSSL_STORE_find() as well as needed tools to construct searches and
1676 get the search data out of them.
1680 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1681 version of OpenSSL should review their configuration settings to ensure
1682 that they are still appropriate for TLSv1.3. For further information see:
1683 <https://wiki.openssl.org/index.php/TLS1.3>
1687 * Grand redesign of the OpenSSL random generator
1689 The default RAND method now utilizes an AES-CTR DRBG according to
1690 NIST standard SP 800-90Ar1. The new random generator is essentially
1691 a port of the default random generator from the OpenSSL FIPS 2.0
1692 object module. It is a hybrid deterministic random bit generator
1693 using an AES-CTR bit stream and which seeds and reseeds itself
1694 automatically using trusted system entropy sources.
1696 Some of its new features are:
1697 - Support for multiple DRBG instances with seed chaining.
1698 - The default RAND method makes use of a DRBG.
1699 - There is a public and private DRBG instance.
1700 - The DRBG instances are fork-safe.
1701 - Keep all global DRBG instances on the secure heap if it is enabled.
1702 - The public and private DRBG instance are per thread for lock free
1705 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1707 * Changed Configure so it only says what it does and doesn't dump
1708 so much data. Instead, ./configdata.pm should be used as a script
1709 to display all sorts of configuration data.
1713 * Added processing of "make variables" to Configure.
1717 * Added SHA512/224 and SHA512/256 algorithm support.
1721 * The last traces of Netware support, first removed in 1.1.0, have
1726 * Get rid of Makefile.shared, and in the process, make the processing
1727 of certain files (rc.obj, or the .def/.map/.opt files produced from
1728 the ordinal files) more visible and hopefully easier to trace and
1729 debug (or make silent).
1733 * Make it possible to have environment variable assignments as
1734 arguments to config / Configure.
1738 * Add multi-prime RSA (RFC 8017) support.
1742 * Add SM3 implemented according to GB/T 32905-2016
1743 *Jack Lloyd <jack.lloyd@ribose.com>,*
1744 *Ronald Tse <ronald.tse@ribose.com>,*
1745 *Erick Borsboom <erick.borsboom@ribose.com>*
1747 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1748 as documented in RFC6066.
1749 Based on a patch from Tomasz Moń
1751 *Filipe Raimundo da Silva*
1753 * Add SM4 implemented according to GB/T 32907-2016.
1754 *Jack Lloyd <jack.lloyd@ribose.com>,*
1755 *Ronald Tse <ronald.tse@ribose.com>,*
1756 *Erick Borsboom <erick.borsboom@ribose.com>*
1758 * Reimplement -newreq-nodes and ERR_error_string_n; the
1759 original author does not agree with the license change.
1763 * Add ARIA AEAD TLS support.
1767 * Some macro definitions to support VS6 have been removed. Visual
1768 Studio 6 has not worked since 1.1.0
1772 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1773 without clearing the errors.
1777 * Add "atfork" functions. If building on a system that without
1778 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1779 requirements. The RAND facility now uses/requires this.
1787 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1788 not possible to disable entirely. However, it's still possible to
1789 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1792 To disable, configure with 'no-ui-console'. 'no-ui' is still
1793 possible to use as an alias. Check at compile time with the
1794 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1795 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1799 * Add a STORE module, which implements a uniform and URI based reader of
1800 stores that can contain keys, certificates, CRLs and numerous other
1801 objects. The main API is loosely based on a few stdio functions,
1802 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1803 OSSL_STORE_error and OSSL_STORE_close.
1804 The implementation uses backends called "loaders" to implement arbitrary
1805 URI schemes. There is one built in "loader" for the 'file' scheme.
1809 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1810 then adjusted to work on FreeBSD 8.4 as well.
1811 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1812 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1816 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1817 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1818 error code calls like this:
1820 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1822 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1823 that can be encoded in C. For the foreseeable future, this will only
1826 *Richard Levitte and Tim Hudson*
1828 * Removed BSD cryptodev engine.
1832 * Add a build target 'build_all_generated', to build all generated files
1833 and only that. This can be used to prepare everything that requires
1834 things like perl for a system that lacks perl and then move everything
1835 to that system and do the rest of the build there.
1839 * In the UI interface, make it possible to duplicate the user data. This
1840 can be used by engines that need to retain the data for a longer time
1841 than just the call where this user data is passed.
1845 * Ignore the '-named_curve auto' value for compatibility of applications
1848 *Tomas Mraz <tmraz@fedoraproject.org>*
1850 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
1851 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
1852 alerts across multiple records (some of which could be empty). In practice
1853 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
1854 prohibits this altogether and other libraries (BoringSSL, NSS) do not
1855 support this at all. Supporting it adds significant complexity to the
1856 record layer, and its removal is unlikely to cause interoperability
1861 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
1862 with Z. These are meant to replace LONG and ZLONG and to be size safe.
1863 The use of LONG and ZLONG is discouraged and scheduled for deprecation
1868 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
1869 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
1871 *Richard Levitte, Andy Polyakov*
1873 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
1878 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1879 platform rather than 'mingw'.
1883 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
1884 success if they are asked to add an object which already exists
1885 in the store. This change cascades to other functions which load
1886 certificates and CRLs.
1890 * x86_64 assembly pack: annotate code with DWARF CFI directives to
1891 facilitate stack unwinding even from assembly subroutines.
1895 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
1896 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
1900 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1901 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1902 which is the minimum version we support.
1906 * Certificate time validation (X509_cmp_time) enforces stricter
1907 compliance with RFC 5280. Fractional seconds and timezone offsets
1908 are no longer allowed.
1912 * Add support for ARIA
1916 * s_client will now send the Server Name Indication (SNI) extension by
1917 default unless the new "-noservername" option is used. The server name is
1918 based on the host provided to the "-connect" option unless overridden by
1919 using "-servername".
1923 * Add support for SipHash
1927 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
1928 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
1929 prevent issues where no progress is being made and the peer continually
1930 sends unrecognised record types, using up resources processing them.
1934 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
1935 using the algorithm defined in
1936 <https://www.akkadia.org/drepper/SHA-crypt.txt>
1940 * Heartbeat support has been removed; the ABI is changed for now.
1942 *Richard Levitte, Rich Salz*
1944 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
1948 * The RSA "null" method, which was partially supported to avoid patent
1949 issues, has been replaced to always returns NULL.
1956 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
1958 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1959 used even when parsing explicit parameters, when loading a serialized key
1960 or calling `EC_GROUP_new_from_ecpkparameters()`/
1961 `EC_GROUP_new_from_ecparameters()`.
1962 This prevents bypass of security hardening and performance gains,
1963 especially for curves with specialized EC_METHODs.
1964 By default, if a key encoded with explicit parameters is loaded and later
1965 serialized, the output is still encoded with explicit parameters, even if
1966 internally a "named" EC_GROUP is used for computation.
1970 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1971 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1972 NULL. After this change, only the cofactor parameter can be NULL. It also
1973 does some minimal sanity checks on the passed order.
1978 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1979 An attack is simple, if the first CMS_recipientInfo is valid but the
1980 second CMS_recipientInfo is chosen ciphertext. If the second
1981 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1982 encryption key will be replaced by garbage, and the message cannot be
1983 decoded, but if the RSA decryption fails, the correct encryption key is
1984 used and the recipient will not notice the attack.
1985 As a work around for this potential attack the length of the decrypted
1986 key must be equal to the cipher default key length, in case the
1987 certifiate is not given and all recipientInfo are tried out.
1988 The old behaviour can be re-enabled in the CMS code by setting the
1989 CMS_DEBUG_DECRYPT flag.
1994 * Use Windows installation paths in the mingw builds
1996 Mingw isn't a POSIX environment per se, which means that Windows
1997 paths should be used for installation.
2002 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2004 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2005 This changes the size when using the genpkey app when no size is given. It
2006 fixes an omission in earlier changes that changed all RSA, DSA and DH
2007 generation apps to use 2048 bits by default.
2011 * Prevent over long nonces in ChaCha20-Poly1305.
2013 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2014 for every encryption operation. RFC 7539 specifies that the nonce value
2015 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2016 and front pads the nonce with 0 bytes if it is less than 12
2017 bytes. However it also incorrectly allows a nonce to be set of up to 16
2018 bytes. In this case only the last 12 bytes are significant and any
2019 additional leading bytes are ignored.
2021 It is a requirement of using this cipher that nonce values are
2022 unique. Messages encrypted using a reused nonce value are susceptible to
2023 serious confidentiality and integrity attacks. If an application changes
2024 the default nonce length to be longer than 12 bytes and then makes a
2025 change to the leading bytes of the nonce expecting the new value to be a
2026 new unique nonce then such an application could inadvertently encrypt
2027 messages with a reused nonce.
2029 Additionally the ignored bytes in a long nonce are not covered by the
2030 integrity guarantee of this cipher. Any application that relies on the
2031 integrity of these ignored leading bytes of a long nonce may be further
2032 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2033 is safe because no such use sets such a long nonce value. However user
2034 applications that use this cipher directly and set a non-default nonce
2035 length to be longer than 12 bytes may be vulnerable.
2037 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2043 * Added SCA hardening for modular field inversion in EC_GROUP through
2044 a new dedicated field_inv() pointer in EC_METHOD.
2045 This also addresses a leakage affecting conversions from projective
2046 to affine coordinates.
2048 *Billy Bob Brumley, Nicola Tuveri*
2050 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2051 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2055 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2059 * Remove the 'dist' target and add a tarball building script. The
2060 'dist' target has fallen out of use, and it shouldn't be
2061 necessary to configure just to create a source distribution.
2065 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2067 * Timing vulnerability in DSA signature generation
2069 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2070 timing side channel attack. An attacker could use variations in the signing
2071 algorithm to recover the private key.
2073 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2078 * Timing vulnerability in ECDSA signature generation
2080 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2081 timing side channel attack. An attacker could use variations in the signing
2082 algorithm to recover the private key.
2084 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2089 * Add coordinate blinding for EC_POINT and implement projective
2090 coordinate blinding for generic prime curves as a countermeasure to
2091 chosen point SCA attacks.
2093 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2095 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2097 * Client DoS due to large DH parameter
2099 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2100 malicious server can send a very large prime value to the client. This will
2101 cause the client to spend an unreasonably long period of time generating a
2102 key for this prime resulting in a hang until the client has finished. This
2103 could be exploited in a Denial Of Service attack.
2105 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2110 * Cache timing vulnerability in RSA Key Generation
2112 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2113 a cache timing side channel attack. An attacker with sufficient access to
2114 mount cache timing attacks during the RSA key generation process could
2115 recover the private key.
2117 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2118 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2123 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2124 parameter is no longer accepted, as it leads to a corrupt table. NULL
2125 pem_str is reserved for alias entries only.
2129 * Revert blinding in ECDSA sign and instead make problematic addition
2130 length-invariant. Switch even to fixed-length Montgomery multiplication.
2134 * Change generating and checking of primes so that the error rate of not
2135 being prime depends on the intended use based on the size of the input.
2136 For larger primes this will result in more rounds of Miller-Rabin.
2137 The maximal error rate for primes with more than 1080 bits is lowered
2140 *Kurt Roeckx, Annie Yousar*
2142 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2146 * Add blinding to ECDSA and DSA signatures to protect against side channel
2147 attacks discovered by Keegan Ryan (NCC Group).
2151 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2152 now allow empty (zero character) pass phrases.
2156 * Certificate time validation (X509_cmp_time) enforces stricter
2157 compliance with RFC 5280. Fractional seconds and timezone offsets
2158 are no longer allowed.
2162 * Fixed a text canonicalisation bug in CMS
2164 Where a CMS detached signature is used with text content the text goes
2165 through a canonicalisation process first prior to signing or verifying a
2166 signature. This process strips trailing space at the end of lines, converts
2167 line terminators to CRLF and removes additional trailing line terminators
2168 at the end of a file. A bug in the canonicalisation process meant that
2169 some characters, such as form-feed, were incorrectly treated as whitespace
2170 and removed. This is contrary to the specification (RFC5485). This fix
2171 could mean that detached text data signed with an earlier version of
2172 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2173 signed with a fixed OpenSSL may fail to verify with an earlier version of
2174 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2175 and use the "-binary" flag (for the "cms" command line application) or set
2176 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2180 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2182 * Constructed ASN.1 types with a recursive definition could exceed the stack
2184 Constructed ASN.1 types with a recursive definition (such as can be found
2185 in PKCS7) could eventually exceed the stack given malicious input with
2186 excessive recursion. This could result in a Denial Of Service attack. There
2187 are no such structures used within SSL/TLS that come from untrusted sources
2188 so this is considered safe.
2190 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2196 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2198 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2199 effectively reduced to only comparing the least significant bit of each
2200 byte. This allows an attacker to forge messages that would be considered as
2201 authenticated in an amount of tries lower than that guaranteed by the
2202 security claims of the scheme. The module can only be compiled by the
2203 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2205 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2211 * Add a build target 'build_all_generated', to build all generated files
2212 and only that. This can be used to prepare everything that requires
2213 things like perl for a system that lacks perl and then move everything
2214 to that system and do the rest of the build there.
2218 * Backport SSL_OP_NO_RENGOTIATION
2220 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2221 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2222 changes this is no longer possible in 1.1.0. Therefore the new
2223 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2224 1.1.0 to provide equivalent functionality.
2226 Note that if an application built against 1.1.0h headers (or above) is run
2227 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2228 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2232 * Removed the OS390-Unix config target. It relied on a script that doesn't
2237 * rsaz_1024_mul_avx2 overflow bug on x86_64
2239 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2240 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2241 Analysis suggests that attacks against RSA and DSA as a result of this
2242 defect would be very difficult to perform and are not believed likely.
2243 Attacks against DH1024 are considered just feasible, because most of the
2244 work necessary to deduce information about a private key may be performed
2245 offline. The amount of resources required for such an attack would be
2246 significant. However, for an attack on TLS to be meaningful, the server
2247 would have to share the DH1024 private key among multiple clients, which is
2248 no longer an option since CVE-2016-0701.
2250 This only affects processors that support the AVX2 but not ADX extensions
2251 like Intel Haswell (4th generation).
2253 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2254 was originally found via the OSS-Fuzz project.
2259 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2261 * bn_sqrx8x_internal carry bug on x86_64
2263 There is a carry propagating bug in the x86_64 Montgomery squaring
2264 procedure. No EC algorithms are affected. Analysis suggests that attacks
2265 against RSA and DSA as a result of this defect would be very difficult to
2266 perform and are not believed likely. Attacks against DH are considered just
2267 feasible (although very difficult) because most of the work necessary to
2268 deduce information about a private key may be performed offline. The amount
2269 of resources required for such an attack would be very significant and
2270 likely only accessible to a limited number of attackers. An attacker would
2271 additionally need online access to an unpatched system using the target
2272 private key in a scenario with persistent DH parameters and a private
2273 key that is shared between multiple clients.
2275 This only affects processors that support the BMI1, BMI2 and ADX extensions
2276 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2278 This issue was reported to OpenSSL by the OSS-Fuzz project.
2283 * Malformed X.509 IPAddressFamily could cause OOB read
2285 If an X.509 certificate has a malformed IPAddressFamily extension,
2286 OpenSSL could do a one-byte buffer overread. The most likely result
2287 would be an erroneous display of the certificate in text format.
2289 This issue was reported to OpenSSL by the OSS-Fuzz project.
2294 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2296 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2297 platform rather than 'mingw'.
2301 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2302 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2303 which is the minimum version we support.
2307 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2309 * Encrypt-Then-Mac renegotiation crash
2311 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2312 negotiated where it was not in the original handshake (or vice-versa) then
2313 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2314 and servers are affected.
2316 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2321 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2323 * Truncated packet could crash via OOB read
2325 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2326 cipher is being used, then a truncated packet can cause that host to
2327 perform an out-of-bounds read, usually resulting in a crash.
2329 This issue was reported to OpenSSL by Robert Święcki of Google.
2334 * Bad (EC)DHE parameters cause a client crash
2336 If a malicious server supplies bad parameters for a DHE or ECDHE key
2337 exchange then this can result in the client attempting to dereference a
2338 NULL pointer leading to a client crash. This could be exploited in a Denial
2341 This issue was reported to OpenSSL by Guido Vranken.
2346 * BN_mod_exp may produce incorrect results on x86_64
2348 There is a carry propagating bug in the x86_64 Montgomery squaring
2349 procedure. No EC algorithms are affected. Analysis suggests that attacks
2350 against RSA and DSA as a result of this defect would be very difficult to
2351 perform and are not believed likely. Attacks against DH are considered just
2352 feasible (although very difficult) because most of the work necessary to
2353 deduce information about a private key may be performed offline. The amount
2354 of resources required for such an attack would be very significant and
2355 likely only accessible to a limited number of attackers. An attacker would
2356 additionally need online access to an unpatched system using the target
2357 private key in a scenario with persistent DH parameters and a private
2358 key that is shared between multiple clients. For example this can occur by
2359 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2360 similar to CVE-2015-3193 but must be treated as a separate problem.
2362 This issue was reported to OpenSSL by the OSS-Fuzz project.
2367 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2369 * ChaCha20/Poly1305 heap-buffer-overflow
2371 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2372 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2373 crash. This issue is not considered to be exploitable beyond a DoS.
2375 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2380 * CMS Null dereference
2382 Applications parsing invalid CMS structures can crash with a NULL pointer
2383 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2384 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2385 structure callback if an attempt is made to free certain invalid encodings.
2386 Only CHOICE structures using a callback which do not handle NULL value are
2389 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2394 * Montgomery multiplication may produce incorrect results
2396 There is a carry propagating bug in the Broadwell-specific Montgomery
2397 multiplication procedure that handles input lengths divisible by, but
2398 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2399 and DH private keys are impossible. This is because the subroutine in
2400 question is not used in operations with the private key itself and an input
2401 of the attacker's direct choice. Otherwise the bug can manifest itself as
2402 transient authentication and key negotiation failures or reproducible
2403 erroneous outcome of public-key operations with specially crafted input.
2404 Among EC algorithms only Brainpool P-512 curves are affected and one
2405 presumably can attack ECDH key negotiation. Impact was not analyzed in
2406 detail, because pre-requisites for attack are considered unlikely. Namely
2407 multiple clients have to choose the curve in question and the server has to
2408 share the private key among them, neither of which is default behaviour.
2409 Even then only clients that chose the curve will be affected.
2411 This issue was publicly reported as transient failures and was not
2412 initially recognized as a security issue. Thanks to Richard Morgan for
2413 providing reproducible case.
2418 * Removed automatic addition of RPATH in shared libraries and executables,
2419 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2423 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2425 * Fix Use After Free for large message sizes
2427 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2428 message larger than approx 16k is received then the underlying buffer to
2429 store the incoming message is reallocated and moved. Unfortunately a
2430 dangling pointer to the old location is left which results in an attempt to
2431 write to the previously freed location. This is likely to result in a
2432 crash, however it could potentially lead to execution of arbitrary code.
2434 This issue only affects OpenSSL 1.1.0a.
2436 This issue was reported to OpenSSL by Robert Święcki.
2441 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2443 * OCSP Status Request extension unbounded memory growth
2445 A malicious client can send an excessively large OCSP Status Request
2446 extension. If that client continually requests renegotiation, sending a
2447 large OCSP Status Request extension each time, then there will be unbounded
2448 memory growth on the server. This will eventually lead to a Denial Of
2449 Service attack through memory exhaustion. Servers with a default
2450 configuration are vulnerable even if they do not support OCSP. Builds using
2451 the "no-ocsp" build time option are not affected.
2453 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2458 * SSL_peek() hang on empty record
2460 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2461 sends an empty record. This could be exploited by a malicious peer in a
2462 Denial Of Service attack.
2464 This issue was reported to OpenSSL by Alex Gaynor.
2469 * Excessive allocation of memory in tls_get_message_header() and
2470 dtls1_preprocess_fragment()
2472 A (D)TLS message includes 3 bytes for its length in the header for the
2473 message. This would allow for messages up to 16Mb in length. Messages of
2474 this length are excessive and OpenSSL includes a check to ensure that a
2475 peer is sending reasonably sized messages in order to avoid too much memory
2476 being consumed to service a connection. A flaw in the logic of version
2477 1.1.0 means that memory for the message is allocated too early, prior to
2478 the excessive message length check. Due to way memory is allocated in
2479 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2480 to service a connection. This could lead to a Denial of Service through
2481 memory exhaustion. However, the excessive message length check still takes
2482 place, and this would cause the connection to immediately fail. Assuming
2483 that the application calls SSL_free() on the failed connection in a timely
2484 manner then the 21Mb of allocated memory will then be immediately freed
2485 again. Therefore the excessive memory allocation will be transitory in
2486 nature. This then means that there is only a security impact if:
2488 1) The application does not call SSL_free() in a timely manner in the event
2489 that the connection fails
2491 2) The application is working in a constrained environment where there is
2492 very little free memory
2494 3) The attacker initiates multiple connection attempts such that there are
2495 multiple connections in a state where memory has been allocated for the
2496 connection; SSL_free() has not yet been called; and there is insufficient
2497 memory to service the multiple requests.
2499 Except in the instance of (1) above any Denial Of Service is likely to be
2500 transitory because as soon as the connection fails the memory is
2501 subsequently freed again in the SSL_free() call. However there is an
2502 increased risk during this period of application crashes due to the lack of
2503 memory - which would then mean a more serious Denial of Service.
2505 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2506 (CVE-2016-6307 and CVE-2016-6308)
2510 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2511 had to be removed. Primary reason is that vendor assembler can't
2512 assemble our modules with -KPIC flag. As result it, assembly
2513 support, was not even available as option. But its lack means
2514 lack of side-channel resistant code, which is incompatible with
2515 security by todays standards. Fortunately gcc is readily available
2516 prepackaged option, which we firmly point at...
2520 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2522 * Windows command-line tool supports UTF-8 opt-in option for arguments
2523 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2524 (to any value) allows Windows user to access PKCS#12 file generated
2525 with Windows CryptoAPI and protected with non-ASCII password, as well
2526 as files generated under UTF-8 locale on Linux also protected with
2531 * To mitigate the SWEET32 attack [CVE-2016-2183][], 3DES cipher suites
2532 have been disabled by default and removed from DEFAULT, just like RC4.
2533 See the RC4 item below to re-enable both.
2537 * The method for finding the storage location for the Windows RAND seed file
2538 has changed. First we check %RANDFILE%. If that is not set then we check
2539 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2540 all else fails we fall back to C:\.
2544 * The EVP_EncryptUpdate() function has had its return type changed from void
2545 to int. A return of 0 indicates and error while a return of 1 indicates
2550 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2551 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2552 off the constant time implementation for RSA, DSA and DH have been made
2553 no-ops and deprecated.
2557 * Windows RAND implementation was simplified to only get entropy by
2558 calling CryptGenRandom(). Various other RAND-related tickets
2561 *Joseph Wylie Yandle, Rich Salz*
2563 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2564 and `OPENSSL_LH_`, respectively. The old names are available
2565 with API compatibility. They new names are now completely documented.
2569 * Unify TYPE_up_ref(obj) methods signature.
2570 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2571 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2572 int (instead of void) like all others TYPE_up_ref() methods.
2573 So now these methods also check the return value of CRYPTO_atomic_add(),
2574 and the validity of object reference counter.
2576 *fdasilvayy@gmail.com*
2578 * With Windows Visual Studio builds, the .pdb files are installed
2579 alongside the installed libraries and executables. For a static
2580 library installation, ossl_static.pdb is the associate compiler
2581 generated .pdb file to be used when linking programs.
2585 * Remove openssl.spec. Packaging files belong with the packagers.
2589 * Automatic Darwin/OSX configuration has had a refresh, it will now
2590 recognise x86_64 architectures automatically. You can still decide
2591 to build for a different bitness with the environment variable
2592 KERNEL_BITS (can be 32 or 64), for example:
2594 KERNEL_BITS=32 ./config
2598 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2599 256 bit AES and HMAC with SHA256.
2603 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2607 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2611 * To enable users to have their own config files and build file templates,
2612 Configure looks in the directory indicated by the environment variable
2613 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2614 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2615 name and is used as is.
2619 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2620 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2621 X509_CERT_FILE_CTX was removed.
2625 * "shared" builds are now the default. To create only static libraries use
2626 the "no-shared" Configure option.
2630 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2631 All of these option have not worked for some while and are fundamental
2636 * Make various cleanup routines no-ops and mark them as deprecated. Most
2637 global cleanup functions are no longer required because they are handled
2638 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2639 Explicitly de-initing can cause problems (e.g. where a library that uses
2640 OpenSSL de-inits, but an application is still using it). The affected
2641 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2642 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2643 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2644 COMP_zlib_cleanup().
2648 * --strict-warnings no longer enables runtime debugging options
2649 such as REF_DEBUG. Instead, debug options are automatically
2650 enabled with '--debug' builds.
2652 *Andy Polyakov, Emilia Käsper*
2654 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2655 have been moved out of the public header files. New functions for managing
2656 these have been added.
2660 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2661 objects have been moved out of the public header files. New
2662 functions for managing these have been added.
2666 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2667 have been moved out of the public header files. New functions for managing
2668 these have been added.
2672 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2673 moved out of the public header files. New functions for managing these
2678 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2682 * Removed the mk1mf build scripts.
2686 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2687 it is always safe to #include a header now.
2691 * Removed the aged BC-32 config and all its supporting scripts
2695 * Removed support for Ultrix, Netware, and OS/2.
2699 * Add support for HKDF.
2701 *Alessandro Ghedini*
2703 * Add support for blake2b and blake2s
2707 * Added support for "pipelining". Ciphers that have the
2708 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2709 encryptions/decryptions simultaneously. There are currently no built-in
2710 ciphers with this property but the expectation is that engines will be able
2711 to offer it to significantly improve throughput. Support has been extended
2712 into libssl so that multiple records for a single connection can be
2713 processed in one go (for >=TLS 1.1).
2717 * Added the AFALG engine. This is an async capable engine which is able to
2718 offload work to the Linux kernel. In this initial version it only supports
2719 AES128-CBC. The kernel must be version 4.1.0 or greater.
2723 * OpenSSL now uses a new threading API. It is no longer necessary to
2724 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2725 are two supported threading models: pthreads and windows threads. It is
2726 also possible to configure OpenSSL at compile time for "no-threads". The
2727 old threading API should no longer be used. The functions have been
2728 replaced with "no-op" compatibility macros.
2730 *Alessandro Ghedini, Matt Caswell*
2732 * Modify behavior of ALPN to invoke callback after SNI/servername
2733 callback, such that updates to the SSL_CTX affect ALPN.
2737 * Add SSL_CIPHER queries for authentication and key-exchange.
2741 * Changes to the DEFAULT cipherlist:
2742 - Prefer (EC)DHE handshakes over plain RSA.
2743 - Prefer AEAD ciphers over legacy ciphers.
2744 - Prefer ECDSA over RSA when both certificates are available.
2745 - Prefer TLSv1.2 ciphers/PRF.
2746 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2751 * Change the ECC default curve list to be this, in order: x25519,
2752 secp256r1, secp521r1, secp384r1.
2756 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2757 disabled by default. They can be re-enabled using the
2758 enable-weak-ssl-ciphers option to Configure.
2762 * If the server has ALPN configured, but supports no protocols that the
2763 client advertises, send a fatal "no_application_protocol" alert.
2764 This behaviour is SHALL in RFC 7301, though it isn't universally
2765 implemented by other servers.
2769 * Add X25519 support.
2770 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2771 for public and private key encoding using the format documented in
2772 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2773 key generation and key derivation.
2775 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2780 * Deprecate SRP_VBASE_get_by_user.
2781 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2782 In order to fix an unavoidable memory leak [CVE-2016-0798][],
2783 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2784 seed, even if the seed is configured.
2786 Users should use SRP_VBASE_get1_by_user instead. Note that in
2787 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2788 also that even though configuring the SRP seed attempts to hide
2789 invalid usernames by continuing the handshake with fake
2790 credentials, this behaviour is not constant time and no strong
2791 guarantees are made that the handshake is indistinguishable from
2792 that of a valid user.
2796 * Configuration change; it's now possible to build dynamic engines
2797 without having to build shared libraries and vice versa. This
2798 only applies to the engines in engines/, those in crypto/engine/
2799 will always be built into libcrypto (i.e. "static").
2801 Building dynamic engines is enabled by default; to disable, use
2802 the configuration option "disable-dynamic-engine".
2804 The only requirements for building dynamic engines are the
2805 presence of the DSO module and building with position independent
2806 code, so they will also automatically be disabled if configuring
2807 with "disable-dso" or "disable-pic".
2809 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2810 are also taken away from openssl/opensslconf.h, as they are
2815 * Configuration change; if there is a known flag to compile
2816 position independent code, it will always be applied on the
2817 libcrypto and libssl object files, and never on the application
2818 object files. This means other libraries that use routines from
2819 libcrypto / libssl can be made into shared libraries regardless
2820 of how OpenSSL was configured.
2822 If this isn't desirable, the configuration options "disable-pic"
2823 or "no-pic" can be used to disable the use of PIC. This will
2824 also disable building shared libraries and dynamic engines.
2828 * Removed JPAKE code. It was experimental and has no wide use.
2832 * The INSTALL_PREFIX Makefile variable has been renamed to
2833 DESTDIR. That makes for less confusion on what this variable
2834 is for. Also, the configuration option --install_prefix is
2839 * Heartbeat for TLS has been removed and is disabled by default
2840 for DTLS; configure with enable-heartbeats. Code that uses the
2841 old #define's might need to be updated.
2843 *Emilia Käsper, Rich Salz*
2845 * Rename REF_CHECK to REF_DEBUG.
2849 * New "unified" build system
2851 The "unified" build system is aimed to be a common system for all
2852 platforms we support. With it comes new support for VMS.
2854 This system builds supports building in a different directory tree
2855 than the source tree. It produces one Makefile (for unix family
2856 or lookalikes), or one descrip.mms (for VMS).
2858 The source of information to make the Makefile / descrip.mms is
2859 small files called 'build.info', holding the necessary
2860 information for each directory with source to compile, and a
2861 template in Configurations, like unix-Makefile.tmpl or
2864 With this change, the library names were also renamed on Windows
2865 and on VMS. They now have names that are closer to the standard
2866 on Unix, and include the major version number, and in certain
2867 cases, the architecture they are built for. See "Notes on shared
2868 libraries" in INSTALL.
2870 We rely heavily on the perl module Text::Template.
2874 * Added support for auto-initialisation and de-initialisation of the library.
2875 OpenSSL no longer requires explicit init or deinit routines to be called,
2876 except in certain circumstances. See the OPENSSL_init_crypto() and
2877 OPENSSL_init_ssl() man pages for further information.
2881 * The arguments to the DTLSv1_listen function have changed. Specifically the
2882 "peer" argument is now expected to be a BIO_ADDR object.
2884 * Rewrite of BIO networking library. The BIO library lacked consistent
2885 support of IPv6, and adding it required some more extensive
2886 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
2887 which hold all types of addresses and chains of address information.
2888 It also introduces a new API, with functions like BIO_socket,
2889 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
2890 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
2891 have been adapted accordingly.
2895 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
2900 * CRIME protection: disable compression by default, even if OpenSSL is
2901 compiled with zlib enabled. Applications can still enable compression
2902 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
2903 using the SSL_CONF library to configure compression.
2907 * The signature of the session callback configured with
2908 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
2909 was explicitly marked as `const unsigned char*` instead of
2914 * Always DPURIFY. Remove the use of uninitialized memory in the
2915 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
2919 * Removed many obsolete configuration items, including
2920 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
2921 MD2_CHAR, MD2_INT, MD2_LONG
2923 IDEA_SHORT, IDEA_LONG
2924 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
2926 *Rich Salz, with advice from Andy Polyakov*
2928 * Many BN internals have been moved to an internal header file.
2930 *Rich Salz with help from Andy Polyakov*
2932 * Configuration and writing out the results from it has changed.
2933 Files such as Makefile include/openssl/opensslconf.h and are now
2934 produced through general templates, such as Makefile.in and
2935 crypto/opensslconf.h.in and some help from the perl module
2938 Also, the center of configuration information is no longer
2939 Makefile. Instead, Configure produces a perl module in
2940 configdata.pm which holds most of the config data (in the hash
2941 table %config), the target data that comes from the target
2942 configuration in one of the `Configurations/*.conf` files (in
2947 * To clarify their intended purposes, the Configure options
2948 --prefix and --openssldir change their semantics, and become more
2949 straightforward and less interdependent.
2951 --prefix shall be used exclusively to give the location INSTALLTOP
2952 where programs, scripts, libraries, include files and manuals are
2953 going to be installed. The default is now /usr/local.
2955 --openssldir shall be used exclusively to give the default
2956 location OPENSSLDIR where certificates, private keys, CRLs are
2957 managed. This is also where the default openssl.cnf gets
2959 If the directory given with this option is a relative path, the
2960 values of both the --prefix value and the --openssldir value will
2961 be combined to become OPENSSLDIR.
2962 The default for --openssldir is INSTALLTOP/ssl.
2964 Anyone who uses --openssldir to specify where OpenSSL is to be
2965 installed MUST change to use --prefix instead.
2969 * The GOST engine was out of date and therefore it has been removed. An up
2970 to date GOST engine is now being maintained in an external repository.
2971 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
2972 support for GOST ciphersuites (these are only activated if a GOST engine
2977 * EGD is no longer supported by default; use enable-egd when
2980 *Ben Kaduk and Rich Salz*
2982 * The distribution now has Makefile.in files, which are used to
2983 create Makefile's when Configure is run. *Configure must be run
2984 before trying to build now.*
2988 * The return value for SSL_CIPHER_description() for error conditions
2993 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
2995 Obtaining and performing DNSSEC validation of TLSA records is
2996 the application's responsibility. The application provides
2997 the TLSA records of its choice to OpenSSL, and these are then
2998 used to authenticate the peer.
3000 The TLSA records need not even come from DNS. They can, for
3001 example, be used to implement local end-entity certificate or
3002 trust-anchor "pinning", where the "pin" data takes the form
3003 of TLSA records, which can augment or replace verification
3004 based on the usual WebPKI public certification authorities.
3008 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3009 continues to support deprecated interfaces in default builds.
3010 However, applications are strongly advised to compile their
3011 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3012 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3013 or the 1.1.0 releases.
3015 In environments in which all applications have been ported to
3016 not use any deprecated interfaces OpenSSL's Configure script
3017 should be used with the --api=1.1.0 option to entirely remove
3018 support for the deprecated features from the library and
3019 unconditionally disable them in the installed headers.
3020 Essentially the same effect can be achieved with the "no-deprecated"
3021 argument to Configure, except that this will always restrict
3022 the build to just the latest API, rather than a fixed API
3025 As applications are ported to future revisions of the API,
3026 they should update their compile-time OPENSSL_API_COMPAT define
3027 accordingly, but in most cases should be able to continue to
3028 compile with later releases.
3030 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3031 0x10000000L and 0x00908000L, respectively. However those
3032 versions did not support the OPENSSL_API_COMPAT feature, and
3033 so applications are not typically tested for explicit support
3034 of just the undeprecated features of either release.
3038 * Add support for setting the minimum and maximum supported protocol.
3039 It can bet set via the SSL_set_min_proto_version() and
3040 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3041 MaxProtocol. It's recommended to use the new APIs to disable
3042 protocols instead of disabling individual protocols using
3043 SSL_set_options() or SSL_CONF's Protocol. This change also
3044 removes support for disabling TLS 1.2 in the OpenSSL TLS
3045 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3049 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3053 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3054 and integrates ECDSA and ECDH functionality into EC. Implementations can
3055 now redirect key generation and no longer need to convert to or from
3058 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3059 include the ec.h header file instead.
3063 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3064 ciphers who are no longer supported and drops support the ephemeral RSA key
3065 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3069 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3070 opaque. For HMAC_CTX, the following constructors and destructors
3073 HMAC_CTX *HMAC_CTX_new(void);
3074 void HMAC_CTX_free(HMAC_CTX *ctx);
3076 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3077 destroy such methods has been added. See EVP_MD_meth_new(3) and
3078 EVP_CIPHER_meth_new(3) for documentation.
3081 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3082 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3083 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3084 an already created structure.
3085 2) For consistency with the majority of our object creators and
3086 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3087 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3088 for deprecated builds.
3092 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3093 cryptographic operations to be performed asynchronously as long as an
3094 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3095 further details. Libssl has also had this capability integrated with the
3096 introduction of the new mode SSL_MODE_ASYNC and associated error
3097 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3098 pages. This work was developed in partnership with Intel Corp.
3102 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3103 always enabled now. If you want to disable the support you should
3104 exclude it using the list of supported ciphers. This also means that the
3105 "-no_ecdhe" option has been removed from s_server.
3109 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3110 SSL_{CTX_}set1_curves() which can set a list.
3114 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3115 curve you want to support using SSL_{CTX_}set1_curves().
3119 * State machine rewrite. The state machine code has been significantly
3120 refactored in order to remove much duplication of code and solve issues
3121 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3122 further details). This change does have some associated API changes.
3123 Notably the SSL_state() function has been removed and replaced by
3124 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3125 SSL_set_state() has been removed altogether. The previous handshake states
3126 defined in ssl.h and ssl3.h have also been removed.
3130 * All instances of the string "ssleay" in the public API were replaced
3131 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3132 Some error codes related to internal RSA_eay API's were renamed.
3136 * The demo files in crypto/threads were moved to demo/threads.
3140 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3143 *Matt Caswell, Rich Salz*
3145 * New ASN.1 embed macro.
3147 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3148 structure is not allocated: it is part of the parent. That is instead of
3156 This reduces memory fragmentation and make it impossible to accidentally
3157 set a mandatory field to NULL.
3159 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3160 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3161 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3166 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3170 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3171 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3172 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3173 DES and RC4 ciphersuites.
3177 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3178 This changes the decoding behaviour for some invalid messages,
3179 though the change is mostly in the more lenient direction, and
3180 legacy behaviour is preserved as much as possible.
3184 * Fix no-stdio build.
3185 *David Woodhouse <David.Woodhouse@intel.com> and also*
3186 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3188 * New testing framework
3189 The testing framework has been largely rewritten and is now using
3190 perl and the perl modules Test::Harness and an extended variant of
3191 Test::More called OpenSSL::Test to do its work. All test scripts in
3192 test/ have been rewritten into test recipes, and all direct calls to
3193 executables in test/Makefile have become individual recipes using the
3194 simplified testing OpenSSL::Test::Simple.
3196 For documentation on our testing modules, do:
3198 perldoc test/testlib/OpenSSL/Test/Simple.pm
3199 perldoc test/testlib/OpenSSL/Test.pm
3203 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3204 are used; the latter aborts on memory leaks (usually checked on exit).
3205 Some undocumented "set malloc, etc., hooks" functions were removed
3206 and others were changed. All are now documented.
3210 * In DSA_generate_parameters_ex, if the provided seed is too short,
3213 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3215 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3216 from RFC4279, RFC4785, RFC5487, RFC5489.
3218 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3219 original RSA_PSK patch.
3223 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3224 era flag was never set throughout the codebase (only read). Also removed
3225 SSL3_FLAGS_POP_BUFFER which was only used if
3226 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3230 * Changed the default name options in the "ca", "crl", "req" and "x509"
3231 to be "oneline" instead of "compat".
3235 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3236 not aware of clients that still exhibit this bug, and the workaround
3237 hasn't been working properly for a while.
3241 * The return type of BIO_number_read() and BIO_number_written() as well as
3242 the corresponding num_read and num_write members in the BIO structure has
3243 changed from unsigned long to uint64_t. On platforms where an unsigned
3244 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3249 * Given the pervasive nature of TLS extensions it is inadvisable to run
3250 OpenSSL without support for them. It also means that maintaining
3251 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3252 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3256 * Removed support for the two export grade static DH ciphersuites
3257 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3258 were newly added (along with a number of other static DH ciphersuites) to
3259 1.0.2. However the two export ones have *never* worked since they were
3260 introduced. It seems strange in any case to be adding new export
3261 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3265 * Version negotiation has been rewritten. In particular SSLv23_method(),
3266 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3267 and turned into macros which simply call the new preferred function names
3268 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3269 should use the new names instead. Also as part of this change the ssl23.h
3270 header file has been removed.
3274 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3275 code and the associated standard is no longer considered fit-for-purpose.
3279 * RT2547 was closed. When generating a private key, try to make the
3280 output file readable only by the owner. This behavior change might
3281 be noticeable when interacting with other software.
3283 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3288 * Added HTTP GET support to the ocsp command.
3292 * Changed default digest for the dgst and enc commands from MD5 to
3297 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3301 * Added support for TLS extended master secret from
3302 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3303 initial patch which was a great help during development.
3307 * All libssl internal structures have been removed from the public header
3308 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3309 now redundant). Users should not attempt to access internal structures
3310 directly. Instead they should use the provided API functions.
3314 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3315 Access to deprecated functions can be re-enabled by running config with
3316 "enable-deprecated". In addition applications wishing to use deprecated
3317 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3318 will, by default, disable some transitive includes that previously existed
3319 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3323 * Added support for OCB mode. OpenSSL has been granted a patent license
3324 compatible with the OpenSSL license for use of OCB. Details are available
3325 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3326 for OCB can be removed by calling config with no-ocb.
3330 * SSLv2 support has been removed. It still supports receiving a SSLv2
3331 compatible client hello.
3335 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3336 done while fixing the error code for the key-too-small case.
3338 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3340 * CA.sh has been removed; use CA.pl instead.
3344 * Removed old DES API.
3348 * Remove various unsupported platforms:
3354 Sinix/ReliantUNIX RM400
3359 16-bit platforms such as WIN16
3363 * Clean up OPENSSL_NO_xxx #define's
3364 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3365 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3366 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3367 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3368 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3369 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3370 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3371 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3372 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3373 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3377 * Cleaned up dead code
3378 Remove all but one '#ifdef undef' which is to be looked at.
3382 * Clean up calling of xxx_free routines.
3383 Just like free(), fix most of the xxx_free routines to accept
3384 NULL. Remove the non-null checks from callers. Save much code.
3388 * Add secure heap for storage of private keys (when possible).
3389 Add BIO_s_secmem(), CBIGNUM, etc.
3390 Contributed by Akamai Technologies under our Corporate CLA.
3394 * Experimental support for a new, fast, unbiased prime candidate generator,
3395 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3397 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3399 * New output format NSS in the sess_id command line tool. This allows
3400 exporting the session id and the master key in NSS keylog format.
3402 *Martin Kaiser <martin@kaiser.cx>*
3404 * Harmonize version and its documentation. -f flag is used to display
3407 *mancha <mancha1@zoho.com>*
3409 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3410 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3412 *mancha <mancha1@zoho.com>*
3414 * Fix some double frees. These are not thought to be exploitable.
3416 *mancha <mancha1@zoho.com>*
3418 * A missing bounds check in the handling of the TLS heartbeat extension
3419 can be used to reveal up to 64k of memory to a connected client or
3422 Thanks for Neel Mehta of Google Security for discovering this bug and to
3423 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3424 preparing the fix [CVE-2014-0160][]
3426 *Adam Langley, Bodo Moeller*
3428 * Fix for the attack described in the paper "Recovering OpenSSL
3429 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3430 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3431 <http://eprint.iacr.org/2014/140>
3433 Thanks to Yuval Yarom and Naomi Benger for discovering this
3434 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
3436 *Yuval Yarom and Naomi Benger*
3438 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3439 this fixes a limitation in previous versions of OpenSSL.
3443 * Experimental encrypt-then-mac support.
3445 Experimental support for encrypt then mac from
3446 draft-gutmann-tls-encrypt-then-mac-02.txt
3448 To enable it set the appropriate extension number (0x42 for the test
3449 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3451 For non-compliant peers (i.e. just about everything) this should have no
3454 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3458 * Add EVP support for key wrapping algorithms, to avoid problems with
3459 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3460 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3461 algorithms and include tests cases.
3465 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3470 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3471 MGF1 digest and OAEP label.
3475 * Make openssl verify return errors.
3477 *Chris Palmer <palmer@google.com> and Ben Laurie*
3479 * New function ASN1_TIME_diff to calculate the difference between two
3480 ASN1_TIME structures or one structure and the current time.
3484 * Update fips_test_suite to support multiple command line options. New
3485 test to induce all self test errors in sequence and check expected
3490 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3491 sign or verify all in one operation.
3495 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3496 test programs and fips_test_suite. Includes functionality to parse
3497 the minimal script output of fipsalgest.pl directly.
3501 * Add authorisation parameter to FIPS_module_mode_set().
3505 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3509 * Use separate DRBG fields for internal and external flags. New function
3510 FIPS_drbg_health_check() to perform on demand health checking. Add
3511 generation tests to fips_test_suite with reduced health check interval to
3512 demonstrate periodic health checking. Add "nodh" option to
3513 fips_test_suite to skip very slow DH test.
3517 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3522 * More extensive health check for DRBG checking many more failure modes.
3523 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3524 combination: call this in fips_test_suite.
3528 * Add support for canonical generation of DSA parameter 'g'. See
3531 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3532 POST to handle HMAC cases.
3536 * Add functions FIPS_module_version() and FIPS_module_version_text()
3537 to return numerical and string versions of the FIPS module number.
3541 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3542 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3543 outside the validated module in the FIPS capable OpenSSL.
3547 * Minor change to DRBG entropy callback semantics. In some cases
3548 there is no multiple of the block length between min_len and
3549 max_len. Allow the callback to return more than max_len bytes
3550 of entropy but discard any extra: it is the callback's responsibility
3551 to ensure that the extra data discarded does not impact the
3552 requested amount of entropy.
3556 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3557 information in FIPS186-3, SP800-57 and SP800-131A.
3561 * CCM support via EVP. Interface is very similar to GCM case except we
3562 must supply all data in one chunk (i.e. no update, final) and the
3563 message length must be supplied if AAD is used. Add algorithm test
3568 * Initial version of POST overhaul. Add POST callback to allow the status
3569 of POST to be monitored and/or failures induced. Modify fips_test_suite
3570 to use callback. Always run all selftests even if one fails.
3574 * XTS support including algorithm test driver in the fips_gcmtest program.
3575 Note: this does increase the maximum key length from 32 to 64 bytes but
3576 there should be no binary compatibility issues as existing applications
3577 will never use XTS mode.
3581 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3582 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3583 performs algorithm blocking for unapproved PRNG types. Also do not
3584 set PRNG type in FIPS_mode_set(): leave this to the application.
3585 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3586 the standard OpenSSL PRNG: set additional data to a date time vector.
3590 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3591 This shouldn't present any incompatibility problems because applications
3592 shouldn't be using these directly and any that are will need to rethink
3593 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3597 * Extensive self tests and health checking required by SP800-90 DRBG.
3598 Remove strength parameter from FIPS_drbg_instantiate and always
3599 instantiate at maximum supported strength.
3603 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3607 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3611 * New function DH_compute_key_padded() to compute a DH key and pad with
3612 leading zeroes if needed: this complies with SP800-56A et al.
3616 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3617 anything, incomplete, subject to change and largely untested at present.
3621 * Modify fipscanisteronly build option to only build the necessary object
3622 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3626 * Add experimental option FIPSSYMS to give all symbols in
3627 fipscanister.o and FIPS or fips prefix. This will avoid
3628 conflicts with future versions of OpenSSL. Add perl script
3629 util/fipsas.pl to preprocess assembly language source files
3630 and rename any affected symbols.
3634 * Add selftest checks and algorithm block of non-fips algorithms in
3635 FIPS mode. Remove DES2 from selftests.
3639 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3640 return internal method without any ENGINE dependencies. Add new
3641 tiny fips sign and verify functions.
3645 * New build option no-ec2m to disable characteristic 2 code.
3649 * New build option "fipscanisteronly". This only builds fipscanister.o
3650 and (currently) associated fips utilities. Uses the file Makefile.fips
3651 instead of Makefile.org as the prototype.
3655 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3656 Update fips_gcmtest to use IV generator.
3660 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3661 setting output buffer to NULL. The `*Final` function must be
3662 called although it will not retrieve any additional data. The tag
3663 can be set or retrieved with a ctrl. The IV length is by default 12
3664 bytes (96 bits) but can be set to an alternative value. If the IV
3665 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3670 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3671 underlying do_cipher function handles all cipher semantics itself
3672 including padding and finalisation. This is useful if (for example)
3673 an ENGINE cipher handles block padding itself. The behaviour of
3674 do_cipher is subtly changed if this flag is set: the return value
3675 is the number of characters written to the output buffer (zero is
3676 no longer an error code) or a negative error code. Also if the
3677 input buffer is NULL and length 0 finalisation should be performed.
3681 * If a candidate issuer certificate is already part of the constructed
3682 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3686 * Improve forward-security support: add functions
3688 void SSL_CTX_set_not_resumable_session_callback(
3689 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3690 void SSL_set_not_resumable_session_callback(
3691 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3693 for use by SSL/TLS servers; the callback function will be called whenever a
3694 new session is created, and gets to decide whether the session may be
3695 cached to make it resumable (return 0) or not (return 1). (As by the
3696 SSL/TLS protocol specifications, the session_id sent by the server will be
3697 empty to indicate that the session is not resumable; also, the server will
3698 not generate RFC 4507 (RFC 5077) session tickets.)
3700 A simple reasonable callback implementation is to return is_forward_secure.
3701 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3702 by the SSL/TLS server library, indicating whether it can provide forward
3705 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3707 * New -verify_name option in command line utilities to set verification
3712 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3713 Add CMAC pkey methods.
3717 * Experimental renegotiation in s_server -www mode. If the client
3718 browses /reneg connection is renegotiated. If /renegcert it is
3719 renegotiated requesting a certificate.
3723 * Add an "external" session cache for debugging purposes to s_server. This
3724 should help trace issues which normally are only apparent in deployed
3725 multi-process servers.
3729 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3730 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3731 BIO_set_cipher() and some obscure PEM functions were changed so they
3732 can now return an error. The RAND changes required a change to the
3733 RAND_METHOD structure.
3737 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3738 a gcc attribute to warn if the result of a function is ignored. This
3739 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3740 whose return value is often ignored.
3744 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3745 These allow SCTs (signed certificate timestamps) to be requested and
3746 validated when establishing a connection.
3748 *Rob Percival <robpercival@google.com>*
3753 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3755 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3756 used even when parsing explicit parameters, when loading a serialized key
3757 or calling `EC_GROUP_new_from_ecpkparameters()`/
3758 `EC_GROUP_new_from_ecparameters()`.
3759 This prevents bypass of security hardening and performance gains,
3760 especially for curves with specialized EC_METHODs.
3761 By default, if a key encoded with explicit parameters is loaded and later
3762 serialized, the output is still encoded with explicit parameters, even if
3763 internally a "named" EC_GROUP is used for computation.
3767 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3768 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3769 NULL. After this change, only the cofactor parameter can be NULL. It also
3770 does some minimal sanity checks on the passed order.
3775 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3776 An attack is simple, if the first CMS_recipientInfo is valid but the
3777 second CMS_recipientInfo is chosen ciphertext. If the second
3778 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3779 encryption key will be replaced by garbage, and the message cannot be
3780 decoded, but if the RSA decryption fails, the correct encryption key is
3781 used and the recipient will not notice the attack.
3782 As a work around for this potential attack the length of the decrypted
3783 key must be equal to the cipher default key length, in case the
3784 certifiate is not given and all recipientInfo are tried out.
3785 The old behaviour can be re-enabled in the CMS code by setting the
3786 CMS_DEBUG_DECRYPT flag.
3791 * Document issue with installation paths in diverse Windows builds
3793 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3794 binaries and run-time config file.
3799 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
3801 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3802 This changes the size when using the genpkey app when no size is given. It
3803 fixes an omission in earlier changes that changed all RSA, DSA and DH
3804 generation apps to use 2048 bits by default.
3808 * Add FIPS support for Android Arm 64-bit
3810 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3811 Module in Version 2.0.10. For some reason, the corresponding target
3812 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3813 built with FIPS support on Android Arm 64-bit. This omission has been
3816 *Matthias St. Pierre*
3818 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
3820 * 0-byte record padding oracle
3822 If an application encounters a fatal protocol error and then calls
3823 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3824 then OpenSSL can respond differently to the calling application if a 0 byte
3825 record is received with invalid padding compared to if a 0 byte record is
3826 received with an invalid MAC. If the application then behaves differently
3827 based on that in a way that is detectable to the remote peer, then this
3828 amounts to a padding oracle that could be used to decrypt data.
3830 In order for this to be exploitable "non-stitched" ciphersuites must be in
3831 use. Stitched ciphersuites are optimised implementations of certain
3832 commonly used ciphersuites. Also the application must call SSL_shutdown()
3833 twice even if a protocol error has occurred (applications should not do
3834 this but some do anyway).
3836 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
3837 Aviram, with additional investigation by Steven Collison and Andrew
3838 Hourselt. It was reported to OpenSSL on 10th December 2018.
3843 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
3847 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
3849 * Microarchitecture timing vulnerability in ECC scalar multiplication
3851 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
3852 shown to be vulnerable to a microarchitecture timing side channel attack.
3853 An attacker with sufficient access to mount local timing attacks during
3854 ECDSA signature generation could recover the private key.
3856 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
3857 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
3863 * Timing vulnerability in DSA signature generation
3865 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
3866 timing side channel attack. An attacker could use variations in the signing
3867 algorithm to recover the private key.
3869 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
3874 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
3875 Module, accidentally introduced while backporting security fixes from the
3876 development branch and hindering the use of ECC in FIPS mode.
3880 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
3882 * Client DoS due to large DH parameter
3884 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
3885 malicious server can send a very large prime value to the client. This will
3886 cause the client to spend an unreasonably long period of time generating a
3887 key for this prime resulting in a hang until the client has finished. This
3888 could be exploited in a Denial Of Service attack.
3890 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
3895 * Cache timing vulnerability in RSA Key Generation
3897 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
3898 a cache timing side channel attack. An attacker with sufficient access to
3899 mount cache timing attacks during the RSA key generation process could
3900 recover the private key.
3902 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
3903 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
3908 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
3909 parameter is no longer accepted, as it leads to a corrupt table. NULL
3910 pem_str is reserved for alias entries only.
3914 * Revert blinding in ECDSA sign and instead make problematic addition
3915 length-invariant. Switch even to fixed-length Montgomery multiplication.
3919 * Change generating and checking of primes so that the error rate of not
3920 being prime depends on the intended use based on the size of the input.
3921 For larger primes this will result in more rounds of Miller-Rabin.
3922 The maximal error rate for primes with more than 1080 bits is lowered
3925 *Kurt Roeckx, Annie Yousar*
3927 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
3931 * Add blinding to ECDSA and DSA signatures to protect against side channel
3932 attacks discovered by Keegan Ryan (NCC Group).
3936 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
3937 now allow empty (zero character) pass phrases.
3941 * Certificate time validation (X509_cmp_time) enforces stricter
3942 compliance with RFC 5280. Fractional seconds and timezone offsets
3943 are no longer allowed.
3947 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
3949 * Constructed ASN.1 types with a recursive definition could exceed the stack
3951 Constructed ASN.1 types with a recursive definition (such as can be found
3952 in PKCS7) could eventually exceed the stack given malicious input with
3953 excessive recursion. This could result in a Denial Of Service attack. There
3954 are no such structures used within SSL/TLS that come from untrusted sources
3955 so this is considered safe.
3957 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
3963 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
3965 * Read/write after SSL object in error state
3967 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
3968 mechanism. The intent was that if a fatal error occurred during a handshake
3969 then OpenSSL would move into the error state and would immediately fail if
3970 you attempted to continue the handshake. This works as designed for the
3971 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
3972 SSL_connect()), however due to a bug it does not work correctly if
3973 SSL_read() or SSL_write() is called directly. In that scenario, if the
3974 handshake fails then a fatal error will be returned in the initial function
3975 call. If SSL_read()/SSL_write() is subsequently called by the application
3976 for the same SSL object then it will succeed and the data is passed without
3977 being decrypted/encrypted directly from the SSL/TLS record layer.
3979 In order to exploit this issue an application bug would have to be present
3980 that resulted in a call to SSL_read()/SSL_write() being issued after having
3981 already received a fatal error.
3983 This issue was reported to OpenSSL by David Benjamin (Google).
3988 * rsaz_1024_mul_avx2 overflow bug on x86_64
3990 There is an overflow bug in the AVX2 Montgomery multiplication procedure
3991 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
3992 Analysis suggests that attacks against RSA and DSA as a result of this
3993 defect would be very difficult to perform and are not believed likely.
3994 Attacks against DH1024 are considered just feasible, because most of the
3995 work necessary to deduce information about a private key may be performed
3996 offline. The amount of resources required for such an attack would be
3997 significant. However, for an attack on TLS to be meaningful, the server
3998 would have to share the DH1024 private key among multiple clients, which is
3999 no longer an option since CVE-2016-0701.
4001 This only affects processors that support the AVX2 but not ADX extensions
4002 like Intel Haswell (4th generation).
4004 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4005 was originally found via the OSS-Fuzz project.
4010 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4012 * bn_sqrx8x_internal carry bug on x86_64
4014 There is a carry propagating bug in the x86_64 Montgomery squaring
4015 procedure. No EC algorithms are affected. Analysis suggests that attacks
4016 against RSA and DSA as a result of this defect would be very difficult to
4017 perform and are not believed likely. Attacks against DH are considered just
4018 feasible (although very difficult) because most of the work necessary to
4019 deduce information about a private key may be performed offline. The amount
4020 of resources required for such an attack would be very significant and
4021 likely only accessible to a limited number of attackers. An attacker would
4022 additionally need online access to an unpatched system using the target
4023 private key in a scenario with persistent DH parameters and a private
4024 key that is shared between multiple clients.
4026 This only affects processors that support the BMI1, BMI2 and ADX extensions
4027 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4029 This issue was reported to OpenSSL by the OSS-Fuzz project.
4034 * Malformed X.509 IPAddressFamily could cause OOB read
4036 If an X.509 certificate has a malformed IPAddressFamily extension,
4037 OpenSSL could do a one-byte buffer overread. The most likely result
4038 would be an erroneous display of the certificate in text format.
4040 This issue was reported to OpenSSL by the OSS-Fuzz project.
4045 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4047 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4048 platform rather than 'mingw'.
4052 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4054 * Truncated packet could crash via OOB read
4056 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4057 cipher is being used, then a truncated packet can cause that host to
4058 perform an out-of-bounds read, usually resulting in a crash.
4060 This issue was reported to OpenSSL by Robert Święcki of Google.
4065 * BN_mod_exp may produce incorrect results on x86_64
4067 There is a carry propagating bug in the x86_64 Montgomery squaring
4068 procedure. No EC algorithms are affected. Analysis suggests that attacks
4069 against RSA and DSA as a result of this defect would be very difficult to
4070 perform and are not believed likely. Attacks against DH are considered just
4071 feasible (although very difficult) because most of the work necessary to
4072 deduce information about a private key may be performed offline. The amount
4073 of resources required for such an attack would be very significant and
4074 likely only accessible to a limited number of attackers. An attacker would
4075 additionally need online access to an unpatched system using the target
4076 private key in a scenario with persistent DH parameters and a private
4077 key that is shared between multiple clients. For example this can occur by
4078 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4079 similar to CVE-2015-3193 but must be treated as a separate problem.
4081 This issue was reported to OpenSSL by the OSS-Fuzz project.
4086 * Montgomery multiplication may produce incorrect results
4088 There is a carry propagating bug in the Broadwell-specific Montgomery
4089 multiplication procedure that handles input lengths divisible by, but
4090 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4091 and DH private keys are impossible. This is because the subroutine in
4092 question is not used in operations with the private key itself and an input
4093 of the attacker's direct choice. Otherwise the bug can manifest itself as
4094 transient authentication and key negotiation failures or reproducible
4095 erroneous outcome of public-key operations with specially crafted input.
4096 Among EC algorithms only Brainpool P-512 curves are affected and one
4097 presumably can attack ECDH key negotiation. Impact was not analyzed in
4098 detail, because pre-requisites for attack are considered unlikely. Namely
4099 multiple clients have to choose the curve in question and the server has to
4100 share the private key among them, neither of which is default behaviour.
4101 Even then only clients that chose the curve will be affected.
4103 This issue was publicly reported as transient failures and was not
4104 initially recognized as a security issue. Thanks to Richard Morgan for
4105 providing reproducible case.
4110 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4111 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4112 prevent issues where no progress is being made and the peer continually
4113 sends unrecognised record types, using up resources processing them.
4117 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4119 * Missing CRL sanity check
4121 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4122 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4123 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4125 This issue only affects the OpenSSL 1.0.2i
4130 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4132 * OCSP Status Request extension unbounded memory growth
4134 A malicious client can send an excessively large OCSP Status Request
4135 extension. If that client continually requests renegotiation, sending a
4136 large OCSP Status Request extension each time, then there will be unbounded
4137 memory growth on the server. This will eventually lead to a Denial Of
4138 Service attack through memory exhaustion. Servers with a default
4139 configuration are vulnerable even if they do not support OCSP. Builds using
4140 the "no-ocsp" build time option are not affected.
4142 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4147 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4150 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4156 * OOB write in MDC2_Update()
4158 An overflow can occur in MDC2_Update() either if called directly or
4159 through the EVP_DigestUpdate() function using MDC2. If an attacker
4160 is able to supply very large amounts of input data after a previous
4161 call to EVP_EncryptUpdate() with a partial block then a length check
4162 can overflow resulting in a heap corruption.
4164 The amount of data needed is comparable to SIZE_MAX which is impractical
4167 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4172 * Malformed SHA512 ticket DoS
4174 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4175 DoS attack where a malformed ticket will result in an OOB read which will
4178 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4179 a custom server callback and ticket lookup mechanism.
4181 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4186 * OOB write in BN_bn2dec()
4188 The function BN_bn2dec() does not check the return value of BN_div_word().
4189 This can cause an OOB write if an application uses this function with an
4190 overly large BIGNUM. This could be a problem if an overly large certificate
4191 or CRL is printed out from an untrusted source. TLS is not affected because
4192 record limits will reject an oversized certificate before it is parsed.
4194 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4199 * OOB read in TS_OBJ_print_bio()
4201 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4202 the total length the OID text representation would use and not the amount
4203 of data written. This will result in OOB reads when large OIDs are
4206 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4211 * Pointer arithmetic undefined behaviour
4213 Avoid some undefined pointer arithmetic
4215 A common idiom in the codebase is to check limits in the following manner:
4218 Where "p" points to some malloc'd data of SIZE bytes and
4221 "len" here could be from some externally supplied data (e.g. from a TLS
4224 The rules of C pointer arithmetic are such that "p + len" is only well
4225 defined where len <= SIZE. Therefore the above idiom is actually
4226 undefined behaviour.
4228 For example this could cause problems if some malloc implementation
4229 provides an address for "p" such that "p + len" actually overflows for
4230 values of len that are too big and therefore p + len < limit.
4232 This issue was reported to OpenSSL by Guido Vranken
4237 * Constant time flag not preserved in DSA signing
4239 Operations in the DSA signing algorithm should run in constant time in
4240 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4241 implementation means that a non-constant time codepath is followed for
4242 certain operations. This has been demonstrated through a cache-timing
4243 attack to be sufficient for an attacker to recover the private DSA key.
4245 This issue was reported by César Pereida (Aalto University), Billy Brumley
4246 (Tampere University of Technology), and Yuval Yarom (The University of
4247 Adelaide and NICTA).
4252 * DTLS buffered message DoS
4254 In a DTLS connection where handshake messages are delivered out-of-order
4255 those messages that OpenSSL is not yet ready to process will be buffered
4256 for later use. Under certain circumstances, a flaw in the logic means that
4257 those messages do not get removed from the buffer even though the handshake
4258 has been completed. An attacker could force up to approx. 15 messages to
4259 remain in the buffer when they are no longer required. These messages will
4260 be cleared when the DTLS connection is closed. The default maximum size for
4261 a message is 100k. Therefore the attacker could force an additional 1500k
4262 to be consumed per connection. By opening many simulataneous connections an
4263 attacker could cause a DoS attack through memory exhaustion.
4265 This issue was reported to OpenSSL by Quan Luo.
4270 * DTLS replay protection DoS
4272 A flaw in the DTLS replay attack protection mechanism means that records
4273 that arrive for future epochs update the replay protection "window" before
4274 the MAC for the record has been validated. This could be exploited by an
4275 attacker by sending a record for the next epoch (which does not have to
4276 decrypt or have a valid MAC), with a very large sequence number. This means
4277 that all subsequent legitimate packets are dropped causing a denial of
4278 service for a specific DTLS connection.
4280 This issue was reported to OpenSSL by the OCAP audit team.
4285 * Certificate message OOB reads
4287 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4288 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4289 theoretical DoS risk but this has not been observed in practice on common
4292 The messages affected are client certificate, client certificate request
4293 and server certificate. As a result the attack can only be performed
4294 against a client or a server which enables client authentication.
4296 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4301 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4303 * Prevent padding oracle in AES-NI CBC MAC check
4305 A MITM attacker can use a padding oracle attack to decrypt traffic
4306 when the connection uses an AES CBC cipher and the server support
4309 This issue was introduced as part of the fix for Lucky 13 padding
4310 attack [CVE-2013-0169][]. The padding check was rewritten to be in
4311 constant time by making sure that always the same bytes are read and
4312 compared against either the MAC or padding bytes. But it no longer
4313 checked that there was enough data to have both the MAC and padding
4316 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4321 * Fix EVP_EncodeUpdate overflow
4323 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4324 Base64 encoding of binary data. If an attacker is able to supply very large
4325 amounts of input data then a length check can overflow resulting in a heap
4328 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4329 the `PEM_write_bio*` family of functions. These are mainly used within the
4330 OpenSSL command line applications, so any application which processes data
4331 from an untrusted source and outputs it as a PEM file should be considered
4332 vulnerable to this issue. User applications that call these APIs directly
4333 with large amounts of untrusted data may also be vulnerable.
4335 This issue was reported by Guido Vranken.
4340 * Fix EVP_EncryptUpdate overflow
4342 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4343 is able to supply very large amounts of input data after a previous call to
4344 EVP_EncryptUpdate() with a partial block then a length check can overflow
4345 resulting in a heap corruption. Following an analysis of all OpenSSL
4346 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4347 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4348 the first called function after an EVP_EncryptInit(), and therefore that
4349 specific call must be safe. The second form is where the length passed to
4350 EVP_EncryptUpdate() can be seen from the code to be some small value and
4351 therefore there is no possibility of an overflow. Since all instances are
4352 one of these two forms, it is believed that there can be no overflows in
4353 internal code due to this problem. It should be noted that
4354 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4355 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4356 of these calls have also been analysed too and it is believed there are no
4357 instances in internal usage where an overflow could occur.
4359 This issue was reported by Guido Vranken.
4364 * Prevent ASN.1 BIO excessive memory allocation
4366 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4367 a short invalid encoding can cause allocation of large amounts of memory
4368 potentially consuming excessive resources or exhausting memory.
4370 Any application parsing untrusted data through d2i BIO functions is
4371 affected. The memory based functions such as d2i_X509() are *not* affected.
4372 Since the memory based functions are used by the TLS library, TLS
4373 applications are not affected.
4375 This issue was reported by Brian Carpenter.
4382 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4383 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4384 in arbitrary stack data being returned in the buffer.
4386 This issue was reported by Guido Vranken.
4391 * Modify behavior of ALPN to invoke callback after SNI/servername
4392 callback, such that updates to the SSL_CTX affect ALPN.
4396 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4401 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4402 methods are enabled and ssl2 is disabled the methods return NULL.
4406 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4408 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4409 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4410 provide any "EXPORT" or "LOW" strength ciphers.
4414 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4415 is by default disabled at build-time. Builds that are not configured with
4416 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4417 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4418 will need to explicitly call either of:
4420 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4422 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4424 as appropriate. Even if either of those is used, or the application
4425 explicitly uses the version-specific SSLv2_method() or its client and
4426 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4427 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4428 ciphers, and SSLv2 56-bit DES are no longer available.
4433 * Fix a double-free in DSA code
4435 A double free bug was discovered when OpenSSL parses malformed DSA private
4436 keys and could lead to a DoS attack or memory corruption for applications
4437 that receive DSA private keys from untrusted sources. This scenario is
4440 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4446 * Disable SRP fake user seed to address a server memory leak.
4448 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4450 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4451 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4452 was changed to ignore the "fake user" SRP seed, even if the seed
4455 Users should use SRP_VBASE_get1_by_user instead. Note that in
4456 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4457 also that even though configuring the SRP seed attempts to hide
4458 invalid usernames by continuing the handshake with fake
4459 credentials, this behaviour is not constant time and no strong
4460 guarantees are made that the handshake is indistinguishable from
4461 that of a valid user.
4466 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4468 In the BN_hex2bn function the number of hex digits is calculated using an
4469 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4470 large values of `i` this can result in `bn_expand` not allocating any
4471 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4472 field as NULL leading to a subsequent NULL ptr deref. For very large values
4473 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4474 In this case memory is allocated to the internal BIGNUM data field, but it
4475 is insufficiently sized leading to heap corruption. A similar issue exists
4476 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4477 is ever called by user applications with very large untrusted hex/dec data.
4478 This is anticipated to be a rare occurrence.
4480 All OpenSSL internal usage of these functions use data that is not expected
4481 to be untrusted, e.g. config file data or application command line
4482 arguments. If user developed applications generate config file data based
4483 on untrusted data then it is possible that this could also lead to security
4484 consequences. This is also anticipated to be rare.
4486 This issue was reported to OpenSSL by Guido Vranken.
4491 * Fix memory issues in `BIO_*printf` functions
4493 The internal `fmtstr` function used in processing a "%s" format string in
4494 the `BIO_*printf` functions could overflow while calculating the length of a
4495 string and cause an OOB read when printing very long strings.
4497 Additionally the internal `doapr_outch` function can attempt to write to an
4498 OOB memory location (at an offset from the NULL pointer) in the event of a
4499 memory allocation failure. In 1.0.2 and below this could be caused where
4500 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4501 could be in processing a very long "%s" format string. Memory leaks can
4504 The first issue may mask the second issue dependent on compiler behaviour.
4505 These problems could enable attacks where large amounts of untrusted data
4506 is passed to the `BIO_*printf` functions. If applications use these functions
4507 in this way then they could be vulnerable. OpenSSL itself uses these
4508 functions when printing out human-readable dumps of ASN.1 data. Therefore
4509 applications that print this data could be vulnerable if the data is from
4510 untrusted sources. OpenSSL command line applications could also be
4511 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4512 as command line arguments.
4514 Libssl is not considered directly vulnerable. Additionally certificates etc
4515 received via remote connections via libssl are also unlikely to be able to
4516 trigger these issues because of message size limits enforced within libssl.
4518 This issue was reported to OpenSSL Guido Vranken.
4523 * Side channel attack on modular exponentiation
4525 A side-channel attack was found which makes use of cache-bank conflicts on
4526 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4527 of RSA keys. The ability to exploit this issue is limited as it relies on
4528 an attacker who has control of code in a thread running on the same
4529 hyper-threaded core as the victim thread which is performing decryptions.
4531 This issue was reported to OpenSSL by Yuval Yarom, The University of
4532 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4533 Nadia Heninger, University of Pennsylvania with more information at
4534 <http://cachebleed.info>.
4539 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4540 if no keysize is specified with default_bits. This fixes an
4541 omission in an earlier change that changed all RSA/DSA key generation
4542 apps to use 2048 bits by default.
4546 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4548 * DH small subgroups
4550 Historically OpenSSL only ever generated DH parameters based on "safe"
4551 primes. More recently (in version 1.0.2) support was provided for
4552 generating X9.42 style parameter files such as those required for RFC 5114
4553 support. The primes used in such files may not be "safe". Where an
4554 application is using DH configured with parameters based on primes that are
4555 not "safe" then an attacker could use this fact to find a peer's private
4556 DH exponent. This attack requires that the attacker complete multiple
4557 handshakes in which the peer uses the same private DH exponent. For example
4558 this could be used to discover a TLS server's private DH exponent if it's
4559 reusing the private DH exponent or it's using a static DH ciphersuite.
4561 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4562 TLS. It is not on by default. If the option is not set then the server
4563 reuses the same private DH exponent for the life of the server process and
4564 would be vulnerable to this attack. It is believed that many popular
4565 applications do set this option and would therefore not be at risk.
4567 The fix for this issue adds an additional check where a "q" parameter is
4568 available (as is the case in X9.42 based parameters). This detects the
4569 only known attack, and is the only possible defense for static DH
4570 ciphersuites. This could have some performance impact.
4572 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4573 default and cannot be disabled. This could have some performance impact.
4575 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4580 * SSLv2 doesn't block disabled ciphers
4582 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4583 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4584 been disabled, provided that the SSLv2 protocol was not also disabled via
4587 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4588 and Sebastian Schinzel.
4593 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4595 * BN_mod_exp may produce incorrect results on x86_64
4597 There is a carry propagating bug in the x86_64 Montgomery squaring
4598 procedure. No EC algorithms are affected. Analysis suggests that attacks
4599 against RSA and DSA as a result of this defect would be very difficult to
4600 perform and are not believed likely. Attacks against DH are considered just
4601 feasible (although very difficult) because most of the work necessary to
4602 deduce information about a private key may be performed offline. The amount
4603 of resources required for such an attack would be very significant and
4604 likely only accessible to a limited number of attackers. An attacker would
4605 additionally need online access to an unpatched system using the target
4606 private key in a scenario with persistent DH parameters and a private
4607 key that is shared between multiple clients. For example this can occur by
4608 default in OpenSSL DHE based SSL/TLS ciphersuites.
4610 This issue was reported to OpenSSL by Hanno Böck.
4615 * Certificate verify crash with missing PSS parameter
4617 The signature verification routines will crash with a NULL pointer
4618 dereference if presented with an ASN.1 signature using the RSA PSS
4619 algorithm and absent mask generation function parameter. Since these
4620 routines are used to verify certificate signature algorithms this can be
4621 used to crash any certificate verification operation and exploited in a
4622 DoS attack. Any application which performs certificate verification is
4623 vulnerable including OpenSSL clients and servers which enable client
4626 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4631 * X509_ATTRIBUTE memory leak
4633 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4634 memory. This structure is used by the PKCS#7 and CMS routines so any
4635 application which reads PKCS#7 or CMS data from untrusted sources is
4636 affected. SSL/TLS is not affected.
4638 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4644 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4645 This changes the decoding behaviour for some invalid messages,
4646 though the change is mostly in the more lenient direction, and
4647 legacy behaviour is preserved as much as possible.
4651 * In DSA_generate_parameters_ex, if the provided seed is too short,
4654 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4656 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4658 * Alternate chains certificate forgery
4660 During certificate verification, OpenSSL will attempt to find an
4661 alternative certificate chain if the first attempt to build such a chain
4662 fails. An error in the implementation of this logic can mean that an
4663 attacker could cause certain checks on untrusted certificates to be
4664 bypassed, such as the CA flag, enabling them to use a valid leaf
4665 certificate to act as a CA and "issue" an invalid certificate.
4667 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4672 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4674 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4675 incompatibility in the handling of HMAC. The previous ABI has now been
4680 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4682 * Malformed ECParameters causes infinite loop
4684 When processing an ECParameters structure OpenSSL enters an infinite loop
4685 if the curve specified is over a specially malformed binary polynomial
4688 This can be used to perform denial of service against any
4689 system which processes public keys, certificate requests or
4690 certificates. This includes TLS clients and TLS servers with
4691 client authentication enabled.
4693 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4698 * Exploitable out-of-bounds read in X509_cmp_time
4700 X509_cmp_time does not properly check the length of the ASN1_TIME
4701 string and can read a few bytes out of bounds. In addition,
4702 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4705 An attacker can use this to craft malformed certificates and CRLs of
4706 various sizes and potentially cause a segmentation fault, resulting in
4707 a DoS on applications that verify certificates or CRLs. TLS clients
4708 that verify CRLs are affected. TLS clients and servers with client
4709 authentication enabled may be affected if they use custom verification
4712 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4713 independently by Hanno Böck.
4718 * PKCS7 crash with missing EnvelopedContent
4720 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4721 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4722 with missing content and trigger a NULL pointer dereference on parsing.
4724 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4725 structures from untrusted sources are affected. OpenSSL clients and
4726 servers are not affected.
4728 This issue was reported to OpenSSL by Michal Zalewski (Google).
4733 * CMS verify infinite loop with unknown hash function
4735 When verifying a signedData message the CMS code can enter an infinite loop
4736 if presented with an unknown hash function OID. This can be used to perform
4737 denial of service against any system which verifies signedData messages using
4739 This issue was reported to OpenSSL by Johannes Bauer.
4744 * Race condition handling NewSessionTicket
4746 If a NewSessionTicket is received by a multi-threaded client when attempting to
4747 reuse a previous ticket then a race condition can occur potentially leading to
4748 a double free of the ticket data.
4753 * Only support 256-bit or stronger elliptic curves with the
4754 'ecdh_auto' setting (server) or by default (client). Of supported
4755 curves, prefer P-256 (both).
4759 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4761 * ClientHello sigalgs DoS fix
4763 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4764 invalid signature algorithms extension a NULL pointer dereference will
4765 occur. This can be exploited in a DoS attack against the server.
4767 This issue was was reported to OpenSSL by David Ramos of Stanford
4771 *Stephen Henson and Matt Caswell*
4773 * Multiblock corrupted pointer fix
4775 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4776 feature only applies on 64 bit x86 architecture platforms that support AES
4777 NI instructions. A defect in the implementation of "multiblock" can cause
4778 OpenSSL's internal write buffer to become incorrectly set to NULL when
4779 using non-blocking IO. Typically, when the user application is using a
4780 socket BIO for writing, this will only result in a failed connection.
4781 However if some other BIO is used then it is likely that a segmentation
4782 fault will be triggered, thus enabling a potential DoS attack.
4784 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4789 * Segmentation fault in DTLSv1_listen fix
4791 The DTLSv1_listen function is intended to be stateless and processes the
4792 initial ClientHello from many peers. It is common for user code to loop
4793 over the call to DTLSv1_listen until a valid ClientHello is received with
4794 an associated cookie. A defect in the implementation of DTLSv1_listen means
4795 that state is preserved in the SSL object from one invocation to the next
4796 that can lead to a segmentation fault. Errors processing the initial
4797 ClientHello can trigger this scenario. An example of such an error could be
4798 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4801 This issue was reported to OpenSSL by Per Allansson.
4806 * Segmentation fault in ASN1_TYPE_cmp fix
4808 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4809 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4810 certificate signature algorithm consistency this can be used to crash any
4811 certificate verification operation and exploited in a DoS attack. Any
4812 application which performs certificate verification is vulnerable including
4813 OpenSSL clients and servers which enable client authentication.
4818 * Segmentation fault for invalid PSS parameters fix
4820 The signature verification routines will crash with a NULL pointer
4821 dereference if presented with an ASN.1 signature using the RSA PSS
4822 algorithm and invalid parameters. Since these routines are used to verify
4823 certificate signature algorithms this can be used to crash any
4824 certificate verification operation and exploited in a DoS attack. Any
4825 application which performs certificate verification is vulnerable including
4826 OpenSSL clients and servers which enable client authentication.
4828 This issue was was reported to OpenSSL by Brian Carpenter.
4833 * ASN.1 structure reuse memory corruption fix
4835 Reusing a structure in ASN.1 parsing may allow an attacker to cause
4836 memory corruption via an invalid write. Such reuse is and has been
4837 strongly discouraged and is believed to be rare.
4839 Applications that parse structures containing CHOICE or ANY DEFINED BY
4840 components may be affected. Certificate parsing (d2i_X509 and related
4841 functions) are however not affected. OpenSSL clients and servers are
4847 * PKCS7 NULL pointer dereferences fix
4849 The PKCS#7 parsing code does not handle missing outer ContentInfo
4850 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
4851 missing content and trigger a NULL pointer dereference on parsing.
4853 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
4854 otherwise parse PKCS#7 structures from untrusted sources are
4855 affected. OpenSSL clients and servers are not affected.
4857 This issue was reported to OpenSSL by Michal Zalewski (Google).
4862 * DoS via reachable assert in SSLv2 servers fix
4864 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
4865 servers that both support SSLv2 and enable export cipher suites by sending
4866 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
4868 This issue was discovered by Sean Burford (Google) and Emilia Käsper
4869 (OpenSSL development team).
4874 * Empty CKE with client auth and DHE fix
4876 If client auth is used then a server can seg fault in the event of a DHE
4877 ciphersuite being selected and a zero length ClientKeyExchange message
4878 being sent by the client. This could be exploited in a DoS attack.
4883 * Handshake with unseeded PRNG fix
4885 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
4886 with an unseeded PRNG. The conditions are:
4887 - The client is on a platform where the PRNG has not been seeded
4888 automatically, and the user has not seeded manually
4889 - A protocol specific client method version has been used (i.e. not
4890 SSL_client_methodv23)
4891 - A ciphersuite is used that does not require additional random data from
4892 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
4894 If the handshake succeeds then the client random that has been used will
4895 have been generated from a PRNG with insufficient entropy and therefore the
4896 output may be predictable.
4898 For example using the following command with an unseeded openssl will
4899 succeed on an unpatched platform:
4901 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
4906 * Use After Free following d2i_ECPrivatekey error fix
4908 A malformed EC private key file consumed via the d2i_ECPrivateKey function
4909 could cause a use after free condition. This, in turn, could cause a double
4910 free in several private key parsing functions (such as d2i_PrivateKey
4911 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
4912 for applications that receive EC private keys from untrusted
4913 sources. This scenario is considered rare.
4915 This issue was discovered by the BoringSSL project and fixed in their
4921 * X509_to_X509_REQ NULL pointer deref fix
4923 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
4924 the certificate key is invalid. This function is rarely used in practice.
4926 This issue was discovered by Brian Carpenter.
4931 * Removed the export ciphers from the DEFAULT ciphers
4935 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
4937 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
4938 ARMv5 through ARMv8, as opposite to "locking" it to single one.
4939 So far those who have to target multiple platforms would compromise
4940 and argue that binary targeting say ARMv5 would still execute on
4941 ARMv8. "Universal" build resolves this compromise by providing
4942 near-optimal performance even on newer platforms.
4946 * Accelerated NIST P-256 elliptic curve implementation for x86_64
4947 (other platforms pending).
4949 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
4951 * Add support for the SignedCertificateTimestampList certificate and
4952 OCSP response extensions from RFC6962.
4956 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
4957 for corner cases. (Certain input points at infinity could lead to
4958 bogus results, with non-infinity inputs mapped to infinity too.)
4962 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
4963 This covers AES, SHA256/512 and GHASH. "Initial" means that most
4964 common cases are optimized and there still is room for further
4965 improvements. Vector Permutation AES for Altivec is also added.
4969 * Add support for little-endian ppc64 Linux target.
4971 *Marcelo Cerri (IBM)*
4973 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
4974 SHA1, SHA256 and GHASH. "Initial" means that most common cases
4975 are optimized and there still is room for further improvements.
4976 Both 32- and 64-bit modes are supported.
4978 *Andy Polyakov, Ard Biesheuvel (Linaro)*
4980 * Improved ARMv7 NEON support.
4984 * Support for SPARC Architecture 2011 crypto extensions, first
4985 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
4986 SHA256/512, MD5, GHASH and modular exponentiation.
4988 *Andy Polyakov, David Miller*
4990 * Accelerated modular exponentiation for Intel processors, a.k.a.
4993 *Shay Gueron & Vlad Krasnov (Intel Corp)*
4995 * Support for new and upcoming Intel processors, including AVX2,
4996 BMI and SHA ISA extensions. This includes additional "stitched"
4997 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5000 This work was sponsored by Intel Corp.
5004 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5005 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5006 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5010 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5011 this fixes a limitation in previous versions of OpenSSL.
5015 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5016 MGF1 digest and OAEP label.
5020 * Add EVP support for key wrapping algorithms, to avoid problems with
5021 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5022 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5023 algorithms and include tests cases.
5027 * Add functions to allocate and set the fields of an ECDSA_METHOD
5030 *Douglas E. Engert, Steve Henson*
5032 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5033 difference in days and seconds between two tm or ASN1_TIME structures.
5037 * Add -rev test option to s_server to just reverse order of characters
5038 received by client and send back to server. Also prints an abbreviated
5039 summary of the connection parameters.
5043 * New option -brief for s_client and s_server to print out a brief summary
5044 of connection parameters.
5048 * Add callbacks for arbitrary TLS extensions.
5050 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5052 * New option -crl_download in several openssl utilities to download CRLs
5053 from CRLDP extension in certificates.
5057 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5061 * New function X509_CRL_diff to generate a delta CRL from the difference
5062 of two full CRLs. Add support to "crl" utility.
5066 * New functions to set lookup_crls function and to retrieve
5067 X509_STORE from X509_STORE_CTX.
5071 * Print out deprecated issuer and subject unique ID fields in
5076 * Extend OCSP I/O functions so they can be used for simple general purpose
5077 HTTP as well as OCSP. New wrapper function which can be used to download
5078 CRLs using the OCSP API.
5082 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5086 * `SSL_CONF*` functions. These provide a common framework for application
5087 configuration using configuration files or command lines.
5091 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5092 message callback and prints the results. Needs compile time option
5093 "enable-ssl-trace". New options to s_client and s_server to enable
5098 * New ctrl and macro to retrieve supported points extensions.
5099 Print out extension in s_server and s_client.
5103 * New functions to retrieve certificate signature and signature
5108 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5113 * New Suite B modes for TLS code. These use and enforce the requirements
5114 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5115 only use Suite B curves. The Suite B modes can be set by using the
5116 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5120 * New chain verification flags for Suite B levels of security. Check
5121 algorithms are acceptable when flags are set in X509_verify_cert.
5125 * Make tls1_check_chain return a set of flags indicating checks passed
5126 by a certificate chain. Add additional tests to handle client
5127 certificates: checks for matching certificate type and issuer name
5132 * If an attempt is made to use a signature algorithm not in the peer
5133 preference list abort the handshake. If client has no suitable
5134 signature algorithms in response to a certificate request do not
5135 use the certificate.
5139 * If server EC tmp key is not in client preference list abort handshake.
5143 * Add support for certificate stores in CERT structure. This makes it
5144 possible to have different stores per SSL structure or one store in
5145 the parent SSL_CTX. Include distinct stores for certificate chain
5146 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5147 to build and store a certificate chain in CERT structure: returning
5148 an error if the chain cannot be built: this will allow applications
5149 to test if a chain is correctly configured.
5151 Note: if the CERT based stores are not set then the parent SSL_CTX
5152 store is used to retain compatibility with existing behaviour.
5156 * New function ssl_set_client_disabled to set a ciphersuite disabled
5157 mask based on the current session, check mask when sending client
5158 hello and checking the requested ciphersuite.
5162 * New ctrls to retrieve and set certificate types in a certificate
5163 request message. Print out received values in s_client. If certificate
5164 types is not set with custom values set sensible values based on
5165 supported signature algorithms.
5169 * Support for distinct client and server supported signature algorithms.
5173 * Add certificate callback. If set this is called whenever a certificate
5174 is required by client or server. An application can decide which
5175 certificate chain to present based on arbitrary criteria: for example
5176 supported signature algorithms. Add very simple example to s_server.
5177 This fixes many of the problems and restrictions of the existing client
5178 certificate callback: for example you can now clear an existing
5179 certificate and specify the whole chain.
5183 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5184 the certificate can be used for (if anything). Set valid_flags field
5185 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5186 to have similar checks in it.
5188 Add new "cert_flags" field to CERT structure and include a "strict mode".
5189 This enforces some TLS certificate requirements (such as only permitting
5190 certificate signature algorithms contained in the supported algorithms
5191 extension) which some implementations ignore: this option should be used
5192 with caution as it could cause interoperability issues.
5196 * Update and tidy signature algorithm extension processing. Work out
5197 shared signature algorithms based on preferences and peer algorithms
5198 and print them out in s_client and s_server. Abort handshake if no
5199 shared signature algorithms.
5203 * Add new functions to allow customised supported signature algorithms
5204 for SSL and SSL_CTX structures. Add options to s_client and s_server
5209 * New function SSL_certs_clear() to delete all references to certificates
5210 from an SSL structure. Before this once a certificate had been added
5211 it couldn't be removed.
5215 * Integrate hostname, email address and IP address checking with certificate
5216 verification. New verify options supporting checking in openssl utility.
5220 * Fixes and wildcard matching support to hostname and email checking
5221 functions. Add manual page.
5223 *Florian Weimer (Red Hat Product Security Team)*
5225 * New functions to check a hostname email or IP address against a
5226 certificate. Add options x509 utility to print results of checks against
5231 * Fix OCSP checking.
5233 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5235 * Initial experimental support for explicitly trusted non-root CAs.
5236 OpenSSL still tries to build a complete chain to a root but if an
5237 intermediate CA has a trust setting included that is used. The first
5238 setting is used: whether to trust (e.g., -addtrust option to the x509
5243 * Add -trusted_first option which attempts to find certificates in the
5244 trusted store even if an untrusted chain is also supplied.
5248 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5249 platform support for Linux and Android.
5253 * Support for linux-x32, ILP32 environment in x86_64 framework.
5257 * Experimental multi-implementation support for FIPS capable OpenSSL.
5258 When in FIPS mode the approved implementations are used as normal,
5259 when not in FIPS mode the internal unapproved versions are used instead.
5260 This means that the FIPS capable OpenSSL isn't forced to use the
5261 (often lower performance) FIPS implementations outside FIPS mode.
5265 * Transparently support X9.42 DH parameters when calling
5266 PEM_read_bio_DHparameters. This means existing applications can handle
5267 the new parameter format automatically.
5271 * Initial experimental support for X9.42 DH parameter format: mainly
5272 to support use of 'q' parameter for RFC5114 parameters.
5276 * Add DH parameters from RFC5114 including test data to dhtest.
5280 * Support for automatic EC temporary key parameter selection. If enabled
5281 the most preferred EC parameters are automatically used instead of
5282 hardcoded fixed parameters. Now a server just has to call:
5283 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5284 support ECDH and use the most appropriate parameters.
5288 * Enhance and tidy EC curve and point format TLS extension code. Use
5289 static structures instead of allocation if default values are used.
5290 New ctrls to set curves we wish to support and to retrieve shared curves.
5291 Print out shared curves in s_server. New options to s_server and s_client
5292 to set list of supported curves.
5296 * New ctrls to retrieve supported signature algorithms and
5297 supported curve values as an array of NIDs. Extend openssl utility
5298 to print out received values.
5302 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5303 between NIDs and the more common NIST names such as "P-256". Enhance
5304 ecparam utility and ECC method to recognise the NIST names for curves.
5308 * Enhance SSL/TLS certificate chain handling to support different
5309 chains for each certificate instead of one chain in the parent SSL_CTX.
5313 * Support for fixed DH ciphersuite client authentication: where both
5314 server and client use DH certificates with common parameters.
5318 * Support for fixed DH ciphersuites: those requiring DH server
5323 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5325 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5326 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5327 X509_CINF_get_signature were reverted post internal team review.
5332 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5334 * OCSP Status Request extension unbounded memory growth
5336 A malicious client can send an excessively large OCSP Status Request
5337 extension. If that client continually requests renegotiation, sending a
5338 large OCSP Status Request extension each time, then there will be unbounded
5339 memory growth on the server. This will eventually lead to a Denial Of
5340 Service attack through memory exhaustion. Servers with a default
5341 configuration are vulnerable even if they do not support OCSP. Builds using
5342 the "no-ocsp" build time option are not affected.
5344 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5349 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5352 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5358 * OOB write in MDC2_Update()
5360 An overflow can occur in MDC2_Update() either if called directly or
5361 through the EVP_DigestUpdate() function using MDC2. If an attacker
5362 is able to supply very large amounts of input data after a previous
5363 call to EVP_EncryptUpdate() with a partial block then a length check
5364 can overflow resulting in a heap corruption.
5366 The amount of data needed is comparable to SIZE_MAX which is impractical
5369 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5374 * Malformed SHA512 ticket DoS
5376 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5377 DoS attack where a malformed ticket will result in an OOB read which will
5380 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5381 a custom server callback and ticket lookup mechanism.
5383 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5388 * OOB write in BN_bn2dec()
5390 The function BN_bn2dec() does not check the return value of BN_div_word().
5391 This can cause an OOB write if an application uses this function with an
5392 overly large BIGNUM. This could be a problem if an overly large certificate
5393 or CRL is printed out from an untrusted source. TLS is not affected because
5394 record limits will reject an oversized certificate before it is parsed.
5396 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5401 * OOB read in TS_OBJ_print_bio()
5403 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5404 the total length the OID text representation would use and not the amount
5405 of data written. This will result in OOB reads when large OIDs are
5408 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5413 * Pointer arithmetic undefined behaviour
5415 Avoid some undefined pointer arithmetic
5417 A common idiom in the codebase is to check limits in the following manner:
5420 Where "p" points to some malloc'd data of SIZE bytes and
5423 "len" here could be from some externally supplied data (e.g. from a TLS
5426 The rules of C pointer arithmetic are such that "p + len" is only well
5427 defined where len <= SIZE. Therefore the above idiom is actually
5428 undefined behaviour.
5430 For example this could cause problems if some malloc implementation
5431 provides an address for "p" such that "p + len" actually overflows for
5432 values of len that are too big and therefore p + len < limit.
5434 This issue was reported to OpenSSL by Guido Vranken
5439 * Constant time flag not preserved in DSA signing
5441 Operations in the DSA signing algorithm should run in constant time in
5442 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5443 implementation means that a non-constant time codepath is followed for
5444 certain operations. This has been demonstrated through a cache-timing
5445 attack to be sufficient for an attacker to recover the private DSA key.
5447 This issue was reported by César Pereida (Aalto University), Billy Brumley
5448 (Tampere University of Technology), and Yuval Yarom (The University of
5449 Adelaide and NICTA).
5454 * DTLS buffered message DoS
5456 In a DTLS connection where handshake messages are delivered out-of-order
5457 those messages that OpenSSL is not yet ready to process will be buffered
5458 for later use. Under certain circumstances, a flaw in the logic means that
5459 those messages do not get removed from the buffer even though the handshake
5460 has been completed. An attacker could force up to approx. 15 messages to
5461 remain in the buffer when they are no longer required. These messages will
5462 be cleared when the DTLS connection is closed. The default maximum size for
5463 a message is 100k. Therefore the attacker could force an additional 1500k
5464 to be consumed per connection. By opening many simulataneous connections an
5465 attacker could cause a DoS attack through memory exhaustion.
5467 This issue was reported to OpenSSL by Quan Luo.
5472 * DTLS replay protection DoS
5474 A flaw in the DTLS replay attack protection mechanism means that records
5475 that arrive for future epochs update the replay protection "window" before
5476 the MAC for the record has been validated. This could be exploited by an
5477 attacker by sending a record for the next epoch (which does not have to
5478 decrypt or have a valid MAC), with a very large sequence number. This means
5479 that all subsequent legitimate packets are dropped causing a denial of
5480 service for a specific DTLS connection.
5482 This issue was reported to OpenSSL by the OCAP audit team.
5487 * Certificate message OOB reads
5489 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5490 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5491 theoretical DoS risk but this has not been observed in practice on common
5494 The messages affected are client certificate, client certificate request
5495 and server certificate. As a result the attack can only be performed
5496 against a client or a server which enables client authentication.
5498 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5503 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5505 * Prevent padding oracle in AES-NI CBC MAC check
5507 A MITM attacker can use a padding oracle attack to decrypt traffic
5508 when the connection uses an AES CBC cipher and the server support
5511 This issue was introduced as part of the fix for Lucky 13 padding
5512 attack [CVE-2013-0169][]. The padding check was rewritten to be in
5513 constant time by making sure that always the same bytes are read and
5514 compared against either the MAC or padding bytes. But it no longer
5515 checked that there was enough data to have both the MAC and padding
5518 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5523 * Fix EVP_EncodeUpdate overflow
5525 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5526 Base64 encoding of binary data. If an attacker is able to supply very large
5527 amounts of input data then a length check can overflow resulting in a heap
5530 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5531 the `PEM_write_bio*` family of functions. These are mainly used within the
5532 OpenSSL command line applications, so any application which processes data
5533 from an untrusted source and outputs it as a PEM file should be considered
5534 vulnerable to this issue. User applications that call these APIs directly
5535 with large amounts of untrusted data may also be vulnerable.
5537 This issue was reported by Guido Vranken.
5542 * Fix EVP_EncryptUpdate overflow
5544 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5545 is able to supply very large amounts of input data after a previous call to
5546 EVP_EncryptUpdate() with a partial block then a length check can overflow
5547 resulting in a heap corruption. Following an analysis of all OpenSSL
5548 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5549 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5550 the first called function after an EVP_EncryptInit(), and therefore that
5551 specific call must be safe. The second form is where the length passed to
5552 EVP_EncryptUpdate() can be seen from the code to be some small value and
5553 therefore there is no possibility of an overflow. Since all instances are
5554 one of these two forms, it is believed that there can be no overflows in
5555 internal code due to this problem. It should be noted that
5556 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5557 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5558 of these calls have also been analysed too and it is believed there are no
5559 instances in internal usage where an overflow could occur.
5561 This issue was reported by Guido Vranken.
5566 * Prevent ASN.1 BIO excessive memory allocation
5568 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5569 a short invalid encoding can casuse allocation of large amounts of memory
5570 potentially consuming excessive resources or exhausting memory.
5572 Any application parsing untrusted data through d2i BIO functions is
5573 affected. The memory based functions such as d2i_X509() are *not* affected.
5574 Since the memory based functions are used by the TLS library, TLS
5575 applications are not affected.
5577 This issue was reported by Brian Carpenter.
5584 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5585 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5586 in arbitrary stack data being returned in the buffer.
5588 This issue was reported by Guido Vranken.
5593 * Modify behavior of ALPN to invoke callback after SNI/servername
5594 callback, such that updates to the SSL_CTX affect ALPN.
5598 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5603 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5604 methods are enabled and ssl2 is disabled the methods return NULL.
5608 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5610 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5611 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5612 provide any "EXPORT" or "LOW" strength ciphers.
5616 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5617 is by default disabled at build-time. Builds that are not configured with
5618 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5619 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5620 will need to explicitly call either of:
5622 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5624 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5626 as appropriate. Even if either of those is used, or the application
5627 explicitly uses the version-specific SSLv2_method() or its client and
5628 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5629 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5630 ciphers, and SSLv2 56-bit DES are no longer available.
5635 * Fix a double-free in DSA code
5637 A double free bug was discovered when OpenSSL parses malformed DSA private
5638 keys and could lead to a DoS attack or memory corruption for applications
5639 that receive DSA private keys from untrusted sources. This scenario is
5642 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5648 * Disable SRP fake user seed to address a server memory leak.
5650 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5652 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5653 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5654 was changed to ignore the "fake user" SRP seed, even if the seed
5657 Users should use SRP_VBASE_get1_by_user instead. Note that in
5658 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5659 also that even though configuring the SRP seed attempts to hide
5660 invalid usernames by continuing the handshake with fake
5661 credentials, this behaviour is not constant time and no strong
5662 guarantees are made that the handshake is indistinguishable from
5663 that of a valid user.
5668 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5670 In the BN_hex2bn function the number of hex digits is calculated using an
5671 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5672 large values of `i` this can result in `bn_expand` not allocating any
5673 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5674 field as NULL leading to a subsequent NULL ptr deref. For very large values
5675 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5676 In this case memory is allocated to the internal BIGNUM data field, but it
5677 is insufficiently sized leading to heap corruption. A similar issue exists
5678 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5679 is ever called by user applications with very large untrusted hex/dec data.
5680 This is anticipated to be a rare occurrence.
5682 All OpenSSL internal usage of these functions use data that is not expected
5683 to be untrusted, e.g. config file data or application command line
5684 arguments. If user developed applications generate config file data based
5685 on untrusted data then it is possible that this could also lead to security
5686 consequences. This is also anticipated to be rare.
5688 This issue was reported to OpenSSL by Guido Vranken.
5693 * Fix memory issues in `BIO_*printf` functions
5695 The internal `fmtstr` function used in processing a "%s" format string in
5696 the `BIO_*printf` functions could overflow while calculating the length of a
5697 string and cause an OOB read when printing very long strings.
5699 Additionally the internal `doapr_outch` function can attempt to write to an
5700 OOB memory location (at an offset from the NULL pointer) in the event of a
5701 memory allocation failure. In 1.0.2 and below this could be caused where
5702 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5703 could be in processing a very long "%s" format string. Memory leaks can
5706 The first issue may mask the second issue dependent on compiler behaviour.
5707 These problems could enable attacks where large amounts of untrusted data
5708 is passed to the `BIO_*printf` functions. If applications use these functions
5709 in this way then they could be vulnerable. OpenSSL itself uses these
5710 functions when printing out human-readable dumps of ASN.1 data. Therefore
5711 applications that print this data could be vulnerable if the data is from
5712 untrusted sources. OpenSSL command line applications could also be
5713 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5714 as command line arguments.
5716 Libssl is not considered directly vulnerable. Additionally certificates etc
5717 received via remote connections via libssl are also unlikely to be able to
5718 trigger these issues because of message size limits enforced within libssl.
5720 This issue was reported to OpenSSL Guido Vranken.
5725 * Side channel attack on modular exponentiation
5727 A side-channel attack was found which makes use of cache-bank conflicts on
5728 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5729 of RSA keys. The ability to exploit this issue is limited as it relies on
5730 an attacker who has control of code in a thread running on the same
5731 hyper-threaded core as the victim thread which is performing decryptions.
5733 This issue was reported to OpenSSL by Yuval Yarom, The University of
5734 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5735 Nadia Heninger, University of Pennsylvania with more information at
5736 <http://cachebleed.info>.
5741 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5742 if no keysize is specified with default_bits. This fixes an
5743 omission in an earlier change that changed all RSA/DSA key generation
5744 apps to use 2048 bits by default.
5748 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5750 * Protection for DH small subgroup attacks
5752 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5753 switched on by default and cannot be disabled. This could have some
5758 * SSLv2 doesn't block disabled ciphers
5760 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5761 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5762 been disabled, provided that the SSLv2 protocol was not also disabled via
5765 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5766 and Sebastian Schinzel.
5771 * Reject DH handshakes with parameters shorter than 1024 bits.
5775 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5777 * Certificate verify crash with missing PSS parameter
5779 The signature verification routines will crash with a NULL pointer
5780 dereference if presented with an ASN.1 signature using the RSA PSS
5781 algorithm and absent mask generation function parameter. Since these
5782 routines are used to verify certificate signature algorithms this can be
5783 used to crash any certificate verification operation and exploited in a
5784 DoS attack. Any application which performs certificate verification is
5785 vulnerable including OpenSSL clients and servers which enable client
5788 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5793 * X509_ATTRIBUTE memory leak
5795 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5796 memory. This structure is used by the PKCS#7 and CMS routines so any
5797 application which reads PKCS#7 or CMS data from untrusted sources is
5798 affected. SSL/TLS is not affected.
5800 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5806 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5807 This changes the decoding behaviour for some invalid messages,
5808 though the change is mostly in the more lenient direction, and
5809 legacy behaviour is preserved as much as possible.
5813 * In DSA_generate_parameters_ex, if the provided seed is too short,
5814 use a random seed, as already documented.
5816 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5818 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
5820 * Alternate chains certificate forgery
5822 During certificate verfification, OpenSSL will attempt to find an
5823 alternative certificate chain if the first attempt to build such a chain
5824 fails. An error in the implementation of this logic can mean that an
5825 attacker could cause certain checks on untrusted certificates to be
5826 bypassed, such as the CA flag, enabling them to use a valid leaf
5827 certificate to act as a CA and "issue" an invalid certificate.
5829 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5835 * Race condition handling PSK identify hint
5837 If PSK identity hints are received by a multi-threaded client then
5838 the values are wrongly updated in the parent SSL_CTX structure. This can
5839 result in a race condition potentially leading to a double free of the
5845 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
5847 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5848 incompatibility in the handling of HMAC. The previous ABI has now been
5851 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
5853 * Malformed ECParameters causes infinite loop
5855 When processing an ECParameters structure OpenSSL enters an infinite loop
5856 if the curve specified is over a specially malformed binary polynomial
5859 This can be used to perform denial of service against any
5860 system which processes public keys, certificate requests or
5861 certificates. This includes TLS clients and TLS servers with
5862 client authentication enabled.
5864 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5869 * Exploitable out-of-bounds read in X509_cmp_time
5871 X509_cmp_time does not properly check the length of the ASN1_TIME
5872 string and can read a few bytes out of bounds. In addition,
5873 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5876 An attacker can use this to craft malformed certificates and CRLs of
5877 various sizes and potentially cause a segmentation fault, resulting in
5878 a DoS on applications that verify certificates or CRLs. TLS clients
5879 that verify CRLs are affected. TLS clients and servers with client
5880 authentication enabled may be affected if they use custom verification
5883 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5884 independently by Hanno Böck.
5889 * PKCS7 crash with missing EnvelopedContent
5891 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5892 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5893 with missing content and trigger a NULL pointer dereference on parsing.
5895 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5896 structures from untrusted sources are affected. OpenSSL clients and
5897 servers are not affected.
5899 This issue was reported to OpenSSL by Michal Zalewski (Google).
5904 * CMS verify infinite loop with unknown hash function
5906 When verifying a signedData message the CMS code can enter an infinite loop
5907 if presented with an unknown hash function OID. This can be used to perform
5908 denial of service against any system which verifies signedData messages using
5910 This issue was reported to OpenSSL by Johannes Bauer.
5915 * Race condition handling NewSessionTicket
5917 If a NewSessionTicket is received by a multi-threaded client when attempting to
5918 reuse a previous ticket then a race condition can occur potentially leading to
5919 a double free of the ticket data.
5924 * Reject DH handshakes with parameters shorter than 768 bits.
5926 *Kurt Roeckx and Emilia Kasper*
5928 * dhparam: generate 2048-bit parameters by default.
5930 *Kurt Roeckx and Emilia Kasper*
5932 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
5934 * Segmentation fault in ASN1_TYPE_cmp fix
5936 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5937 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5938 certificate signature algorithm consistency this can be used to crash any
5939 certificate verification operation and exploited in a DoS attack. Any
5940 application which performs certificate verification is vulnerable including
5941 OpenSSL clients and servers which enable client authentication.
5946 * ASN.1 structure reuse memory corruption fix
5948 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5949 memory corruption via an invalid write. Such reuse is and has been
5950 strongly discouraged and is believed to be rare.
5952 Applications that parse structures containing CHOICE or ANY DEFINED BY
5953 components may be affected. Certificate parsing (d2i_X509 and related
5954 functions) are however not affected. OpenSSL clients and servers are
5960 * PKCS7 NULL pointer dereferences fix
5962 The PKCS#7 parsing code does not handle missing outer ContentInfo
5963 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5964 missing content and trigger a NULL pointer dereference on parsing.
5966 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5967 otherwise parse PKCS#7 structures from untrusted sources are
5968 affected. OpenSSL clients and servers are not affected.
5970 This issue was reported to OpenSSL by Michal Zalewski (Google).
5975 * DoS via reachable assert in SSLv2 servers fix
5977 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5978 servers that both support SSLv2 and enable export cipher suites by sending
5979 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5981 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5982 (OpenSSL development team).
5987 * Use After Free following d2i_ECPrivatekey error fix
5989 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5990 could cause a use after free condition. This, in turn, could cause a double
5991 free in several private key parsing functions (such as d2i_PrivateKey
5992 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5993 for applications that receive EC private keys from untrusted
5994 sources. This scenario is considered rare.
5996 This issue was discovered by the BoringSSL project and fixed in their
6002 * X509_to_X509_REQ NULL pointer deref fix
6004 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6005 the certificate key is invalid. This function is rarely used in practice.
6007 This issue was discovered by Brian Carpenter.
6012 * Removed the export ciphers from the DEFAULT ciphers
6016 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6018 * Build fixes for the Windows and OpenVMS platforms
6020 *Matt Caswell and Richard Levitte*
6022 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6024 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6025 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6026 dereference. This could lead to a Denial Of Service attack. Thanks to
6027 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6032 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6033 dtls1_buffer_record function under certain conditions. In particular this
6034 could occur if an attacker sent repeated DTLS records with the same
6035 sequence number but for the next epoch. The memory leak could be exploited
6036 by an attacker in a Denial of Service attack through memory exhaustion.
6037 Thanks to Chris Mueller for reporting this issue.
6042 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6043 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6044 method would be set to NULL which could later result in a NULL pointer
6045 dereference. Thanks to Frank Schmirler for reporting this issue.
6050 * Abort handshake if server key exchange message is omitted for ephemeral
6053 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6054 reporting this issue.
6059 * Remove non-export ephemeral RSA code on client and server. This code
6060 violated the TLS standard by allowing the use of temporary RSA keys in
6061 non-export ciphersuites and could be used by a server to effectively
6062 downgrade the RSA key length used to a value smaller than the server
6063 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6064 INRIA or reporting this issue.
6069 * Fixed issue where DH client certificates are accepted without verification.
6070 An OpenSSL server will accept a DH certificate for client authentication
6071 without the certificate verify message. This effectively allows a client to
6072 authenticate without the use of a private key. This only affects servers
6073 which trust a client certificate authority which issues certificates
6074 containing DH keys: these are extremely rare and hardly ever encountered.
6075 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6081 * Ensure that the session ID context of an SSL is updated when its
6082 SSL_CTX is updated via SSL_set_SSL_CTX.
6084 The session ID context is typically set from the parent SSL_CTX,
6085 and can vary with the CTX.
6089 * Fix various certificate fingerprint issues.
6091 By using non-DER or invalid encodings outside the signed portion of a
6092 certificate the fingerprint can be changed without breaking the signature.
6093 Although no details of the signed portion of the certificate can be changed
6094 this can cause problems with some applications: e.g. those using the
6095 certificate fingerprint for blacklists.
6097 1. Reject signatures with non zero unused bits.
6099 If the BIT STRING containing the signature has non zero unused bits reject
6100 the signature. All current signature algorithms require zero unused bits.
6102 2. Check certificate algorithm consistency.
6104 Check the AlgorithmIdentifier inside TBS matches the one in the
6105 certificate signature. NB: this will result in signature failure
6106 errors for some broken certificates.
6108 Thanks to Konrad Kraszewski from Google for reporting this issue.
6110 3. Check DSA/ECDSA signatures use DER.
6112 Re-encode DSA/ECDSA signatures and compare with the original received
6113 signature. Return an error if there is a mismatch.
6115 This will reject various cases including garbage after signature
6116 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6117 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6118 (negative or with leading zeroes).
6120 Further analysis was conducted and fixes were developed by Stephen Henson
6121 of the OpenSSL core team.
6127 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6128 results on some platforms, including x86_64. This bug occurs at random
6129 with a very low probability, and is not known to be exploitable in any
6130 way, though its exact impact is difficult to determine. Thanks to Pieter
6131 Wuille (Blockstream) who reported this issue and also suggested an initial
6132 fix. Further analysis was conducted by the OpenSSL development team and
6133 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6134 the OpenSSL core team.
6139 * Do not resume sessions on the server if the negotiated protocol
6140 version does not match the session's version. Resuming with a different
6141 version, while not strictly forbidden by the RFC, is of questionable
6142 sanity and breaks all known clients.
6144 *David Benjamin, Emilia Käsper*
6146 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6147 early CCS messages during renegotiation. (Note that because
6148 renegotiation is encrypted, this early CCS was not exploitable.)
6152 * Tighten client-side session ticket handling during renegotiation:
6153 ensure that the client only accepts a session ticket if the server sends
6154 the extension anew in the ServerHello. Previously, a TLS client would
6155 reuse the old extension state and thus accept a session ticket if one was
6156 announced in the initial ServerHello.
6158 Similarly, ensure that the client requires a session ticket if one
6159 was advertised in the ServerHello. Previously, a TLS client would
6160 ignore a missing NewSessionTicket message.
6164 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6168 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6169 sends a carefully crafted handshake message, to cause OpenSSL to fail
6170 to free up to 64k of memory causing a memory leak. This could be
6171 exploited in a Denial Of Service attack. This issue affects OpenSSL
6172 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6173 whether SRTP is used or configured. Implementations of OpenSSL that
6174 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6176 The fix was developed by the OpenSSL team.
6181 * Session Ticket Memory Leak.
6183 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6184 integrity of that ticket is first verified. In the event of a session
6185 ticket integrity check failing, OpenSSL will fail to free memory
6186 causing a memory leak. By sending a large number of invalid session
6187 tickets an attacker could exploit this issue in a Denial Of Service
6193 * Build option no-ssl3 is incomplete.
6195 When OpenSSL is configured with "no-ssl3" as a build option, servers
6196 could accept and complete a SSL 3.0 handshake, and clients could be
6197 configured to send them.
6200 *Akamai and the OpenSSL team*
6202 * Add support for TLS_FALLBACK_SCSV.
6203 Client applications doing fallback retries should call
6204 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6207 *Adam Langley, Bodo Moeller*
6209 * Add additional DigestInfo checks.
6211 Re-encode DigestInto in DER and check against the original when
6212 verifying RSA signature: this will reject any improperly encoded
6213 DigestInfo structures.
6215 Note: this is a precautionary measure and no attacks are currently known.
6219 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6221 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6222 SRP code can be overrun an internal buffer. Add sanity check that
6223 g, A, B < N to SRP code.
6225 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6226 Group for discovering this issue.
6231 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6232 TLS 1.0 instead of higher protocol versions when the ClientHello message
6233 is badly fragmented. This allows a man-in-the-middle attacker to force a
6234 downgrade to TLS 1.0 even if both the server and the client support a
6235 higher protocol version, by modifying the client's TLS records.
6237 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6238 researching this issue.
6243 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6244 to a denial of service attack. A malicious server can crash the client
6245 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6246 ciphersuite and sending carefully crafted handshake messages.
6248 Thanks to Felix Gröbert (Google) for discovering and researching this
6254 * By sending carefully crafted DTLS packets an attacker could cause openssl
6255 to leak memory. This can be exploited through a Denial of Service attack.
6256 Thanks to Adam Langley for discovering and researching this issue.
6261 * An attacker can force openssl to consume large amounts of memory whilst
6262 processing DTLS handshake messages. This can be exploited through a
6263 Denial of Service attack.
6264 Thanks to Adam Langley for discovering and researching this issue.
6269 * An attacker can force an error condition which causes openssl to crash
6270 whilst processing DTLS packets due to memory being freed twice. This
6271 can be exploited through a Denial of Service attack.
6272 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6278 * If a multithreaded client connects to a malicious server using a resumed
6279 session and the server sends an ec point format extension it could write
6280 up to 255 bytes to freed memory.
6282 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6288 * A malicious server can crash an OpenSSL client with a null pointer
6289 dereference (read) by specifying an SRP ciphersuite even though it was not
6290 properly negotiated with the client. This can be exploited through a
6291 Denial of Service attack.
6293 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6294 discovering and researching this issue.
6299 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6300 X509_name_oneline, X509_name_print_ex et al. to leak some information
6301 from the stack. Applications may be affected if they echo pretty printing
6302 output to the attacker.
6304 Thanks to Ivan Fratric (Google) for discovering this issue.
6307 *Emilia Käsper, and Steve Henson*
6309 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6310 for corner cases. (Certain input points at infinity could lead to
6311 bogus results, with non-infinity inputs mapped to infinity too.)
6315 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6317 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6318 handshake can force the use of weak keying material in OpenSSL
6319 SSL/TLS clients and servers.
6321 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6322 researching this issue. [CVE-2014-0224][]
6324 *KIKUCHI Masashi, Steve Henson*
6326 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6327 OpenSSL DTLS client the code can be made to recurse eventually crashing
6330 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6333 *Imre Rad, Steve Henson*
6335 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6336 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6337 client or server. This is potentially exploitable to run arbitrary
6338 code on a vulnerable client or server.
6340 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
6342 *Jüri Aedla, Steve Henson*
6344 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6345 are subject to a denial of service attack.
6347 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6348 this issue. [CVE-2014-3470][]
6350 *Felix Gröbert, Ivan Fratric, Steve Henson*
6352 * Harmonize version and its documentation. -f flag is used to display
6355 *mancha <mancha1@zoho.com>*
6357 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6358 in i2d_ECPrivateKey.
6360 *mancha <mancha1@zoho.com>*
6362 * Fix some double frees. These are not thought to be exploitable.
6364 *mancha <mancha1@zoho.com>*
6366 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6368 * A missing bounds check in the handling of the TLS heartbeat extension
6369 can be used to reveal up to 64k of memory to a connected client or
6372 Thanks for Neel Mehta of Google Security for discovering this bug and to
6373 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6374 preparing the fix [CVE-2014-0160][]
6376 *Adam Langley, Bodo Moeller*
6378 * Fix for the attack described in the paper "Recovering OpenSSL
6379 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6380 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6381 <http://eprint.iacr.org/2014/140>
6383 Thanks to Yuval Yarom and Naomi Benger for discovering this
6384 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
6386 *Yuval Yarom and Naomi Benger*
6388 * TLS pad extension: draft-agl-tls-padding-03
6390 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6391 TLS client Hello record length value would otherwise be > 255 and
6392 less that 512 pad with a dummy extension containing zeroes so it
6393 is at least 512 bytes long.
6395 *Adam Langley, Steve Henson*
6397 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6399 * Fix for TLS record tampering bug. A carefully crafted invalid
6400 handshake could crash OpenSSL with a NULL pointer exception.
6401 Thanks to Anton Johansson for reporting this issues.
6404 * Keep original DTLS digest and encryption contexts in retransmission
6405 structures so we can use the previous session parameters if they need
6406 to be resent. [CVE-2013-6450][]
6410 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6411 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6412 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6413 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6414 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6415 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6417 *Rob Stradling, Adam Langley*
6419 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6421 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6422 supporting platforms or when small records were transferred.
6424 *Andy Polyakov, Steve Henson*
6426 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6428 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6430 This addresses the flaw in CBC record processing discovered by
6431 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6432 at: <http://www.isg.rhul.ac.uk/tls/>
6434 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6435 Security Group at Royal Holloway, University of London
6436 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6437 Emilia Käsper for the initial patch.
6440 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6442 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6443 ciphersuites which can be exploited in a denial of service attack.
6444 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6445 and detecting this bug and to Wolfgang Ettlinger
6446 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6451 * Return an error when checking OCSP signatures when key is NULL.
6452 This fixes a DoS attack. [CVE-2013-0166][]
6456 * Make openssl verify return errors.
6458 *Chris Palmer <palmer@google.com> and Ben Laurie*
6460 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6461 the right response is stapled. Also change SSL_get_certificate()
6462 so it returns the certificate actually sent.
6463 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6465 *Rob Stradling <rob.stradling@comodo.com>*
6467 * Fix possible deadlock when decoding public keys.
6471 * Don't use TLS 1.0 record version number in initial client hello
6476 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6478 * Sanity check record length before skipping explicit IV in TLS
6479 1.2, 1.1 and DTLS to fix DoS attack.
6481 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6482 fuzzing as a service testing platform.
6487 * Initialise tkeylen properly when encrypting CMS messages.
6488 Thanks to Solar Designer of Openwall for reporting this issue.
6492 * In FIPS mode don't try to use composite ciphers as they are not
6497 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6499 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6500 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6501 mean any application compiled against OpenSSL 1.0.0 headers setting
6502 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6503 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6504 0x10000000L Any application which was previously compiled against
6505 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6506 will need to be recompiled as a result. Letting be results in
6507 inability to disable specifically TLS 1.1 and in client context,
6508 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6512 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6513 disable just protocol X, but all protocols above X *if* there are
6514 protocols *below* X still enabled. In more practical terms it means
6515 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6516 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6517 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6522 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6524 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6525 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6526 in CRYPTO_realloc_clean.
6528 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6529 issue and to Adam Langley <agl@chromium.org> for fixing it.
6532 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6534 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6538 * Workarounds for some broken servers that "hang" if a client hello
6539 record length exceeds 255 bytes.
6541 1. Do not use record version number > TLS 1.0 in initial client
6542 hello: some (but not all) hanging servers will now work.
6543 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6544 the number of ciphers sent in the client hello. This should be
6545 set to an even number, such as 50, for example by passing:
6546 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6547 Most broken servers should now work.
6548 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6549 TLS 1.2 client support entirely.
6553 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6557 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6559 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6560 STRING form instead of a DigestInfo.
6564 * The format used for MDC2 RSA signatures is inconsistent between EVP
6565 and the RSA_sign/RSA_verify functions. This was made more apparent when
6566 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6567 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6568 the correct format in RSA_verify so both forms transparently work.
6572 * Some servers which support TLS 1.0 can choke if we initially indicate
6573 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6574 encrypted premaster secret. As a workaround use the maximum permitted
6575 client version in client hello, this should keep such servers happy
6576 and still work with previous versions of OpenSSL.
6580 * Add support for TLS/DTLS heartbeats.
6582 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6584 * Add support for SCTP.
6586 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6588 * Improved PRNG seeding for VOS.
6590 *Paul Green <Paul.Green@stratus.com>*
6592 * Extensive assembler packs updates, most notably:
6594 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6595 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6596 - x86_64: bit-sliced AES implementation;
6597 - ARM: NEON support, contemporary platforms optimizations;
6598 - s390x: z196 support;
6599 - `*`: GHASH and GF(2^m) multiplication implementations;
6603 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6604 (removal of unnecessary code)
6606 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6608 * Add TLS key material exporter from RFC 5705.
6612 * Add DTLS-SRTP negotiation from RFC 5764.
6616 * Add Next Protocol Negotiation,
6617 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6618 disabled with a no-npn flag to config or Configure. Code donated
6621 *Adam Langley <agl@google.com> and Ben Laurie*
6623 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6624 NIST-P256, NIST-P521, with constant-time single point multiplication on
6625 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6626 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6627 Code made available under Apache License version 2.0.
6629 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6630 line to include this in your build of OpenSSL, and run "make depend" (or
6631 "make update"). This enables the following EC_METHODs:
6633 EC_GFp_nistp224_method()
6634 EC_GFp_nistp256_method()
6635 EC_GFp_nistp521_method()
6637 EC_GROUP_new_by_curve_name() will automatically use these (while
6638 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6641 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6643 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6644 all platforms. Move ssize_t definition from e_os.h to the public
6645 header file e_os2.h as it now appears in public header file cms.h
6649 * New -sigopt option to the ca, req and x509 utilities. Additional
6650 signature parameters can be passed using this option and in
6655 * Add RSA PSS signing function. This will generate and set the
6656 appropriate AlgorithmIdentifiers for PSS based on those in the
6657 corresponding EVP_MD_CTX structure. No application support yet.
6661 * Support for companion algorithm specific ASN1 signing routines.
6662 New function ASN1_item_sign_ctx() signs a pre-initialised
6663 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6664 the appropriate parameters.
6668 * Add new algorithm specific ASN1 verification initialisation function
6669 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6670 handling will be the same no matter what EVP_PKEY_METHOD is used.
6671 Add a PSS handler to support verification of PSS signatures: checked
6672 against a number of sample certificates.
6676 * Add signature printing for PSS. Add PSS OIDs.
6678 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6680 * Add algorithm specific signature printing. An individual ASN1 method
6681 can now print out signatures instead of the standard hex dump.
6683 More complex signatures (e.g. PSS) can print out more meaningful
6684 information. Include DSA version that prints out the signature
6689 * Password based recipient info support for CMS library: implementing
6694 * Split password based encryption into PBES2 and PBKDF2 functions. This
6695 neatly separates the code into cipher and PBE sections and is required
6696 for some algorithms that split PBES2 into separate pieces (such as
6697 password based CMS).
6701 * Session-handling fixes:
6702 - Fix handling of connections that are resuming with a session ID,
6703 but also support Session Tickets.
6704 - Fix a bug that suppressed issuing of a new ticket if the client
6705 presented a ticket with an expired session.
6706 - Try to set the ticket lifetime hint to something reasonable.
6707 - Make tickets shorter by excluding irrelevant information.
6708 - On the client side, don't ignore renewed tickets.
6710 *Adam Langley, Bodo Moeller (Google)*
6712 * Fix PSK session representation.
6716 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6718 This work was sponsored by Intel.
6722 * Add GCM support to TLS library. Some custom code is needed to split
6723 the IV between the fixed (from PRF) and explicit (from TLS record)
6724 portions. This adds all GCM ciphersuites supported by RFC5288 and
6725 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6726 add a special AESGCM string for GCM only.
6730 * Expand range of ctrls for AES GCM. Permit setting invocation
6731 field on decrypt and retrieval of invocation field only on encrypt.
6735 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6736 As required by RFC5289 these ciphersuites cannot be used if for
6737 versions of TLS earlier than 1.2.
6741 * For FIPS capable OpenSSL interpret a NULL default public key method
6742 as unset and return the appropriate default but do *not* set the default.
6743 This means we can return the appropriate method in applications that
6744 switch between FIPS and non-FIPS modes.
6748 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6749 ENGINE is used then we cannot handle that in the FIPS module so we
6750 keep original code iff non-FIPS operations are allowed.
6754 * Add -attime option to openssl utilities.
6756 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6758 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6762 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6763 FIPS EC methods unconditionally for now.
6767 * New build option no-ec2m to disable characteristic 2 code.
6771 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6772 all cases can be covered as some introduce binary incompatibilities.
6776 * Redirect RSA operations to FIPS module including keygen,
6777 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6781 * Add similar low level API blocking to ciphers.
6785 * Low level digest APIs are not approved in FIPS mode: any attempt
6786 to use these will cause a fatal error. Applications that *really* want
6787 to use them can use the `private_*` version instead.
6791 * Redirect cipher operations to FIPS module for FIPS builds.
6795 * Redirect digest operations to FIPS module for FIPS builds.
6799 * Update build system to add "fips" flag which will link in fipscanister.o
6800 for static and shared library builds embedding a signature if needed.
6804 * Output TLS supported curves in preference order instead of numerical
6805 order. This is currently hardcoded for the highest order curves first.
6806 This should be configurable so applications can judge speed vs strength.
6810 * Add TLS v1.2 server support for client authentication.
6814 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6819 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6820 FIPS modules versions.
6824 * Add TLS v1.2 client side support for client authentication. Keep cache
6825 of handshake records longer as we don't know the hash algorithm to use
6826 until after the certificate request message is received.
6830 * Initial TLS v1.2 client support. Add a default signature algorithms
6831 extension including all the algorithms we support. Parse new signature
6832 format in client key exchange. Relax some ECC signing restrictions for
6833 TLS v1.2 as indicated in RFC5246.
6837 * Add server support for TLS v1.2 signature algorithms extension. Switch
6838 to new signature format when needed using client digest preference.
6839 All server ciphersuites should now work correctly in TLS v1.2. No client
6840 support yet and no support for client certificates.
6844 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
6845 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
6846 ciphersuites. At present only RSA key exchange ciphersuites work with
6847 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
6848 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
6849 and version checking.
6853 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
6854 with this defined it will not be affected by any changes to ssl internal
6855 structures. Add several utility functions to allow openssl application
6856 to work with OPENSSL_NO_SSL_INTERN defined.
6860 * A long standing patch to add support for SRP from EdelWeb (Peter
6861 Sylvester and Christophe Renou) was integrated.
6862 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
6863 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
6866 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
6870 * Permit abbreviated handshakes when renegotiating using the function
6871 SSL_renegotiate_abbreviated().
6873 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6875 * Add call to ENGINE_register_all_complete() to
6876 ENGINE_load_builtin_engines(), so some implementations get used
6877 automatically instead of needing explicit application support.
6881 * Add support for TLS key exporter as described in RFC5705.
6883 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
6885 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
6886 a few changes are required:
6888 Add SSL_OP_NO_TLSv1_1 flag.
6889 Add TLSv1_1 methods.
6890 Update version checking logic to handle version 1.1.
6891 Add explicit IV handling (ported from DTLS code).
6892 Add command line options to s_client/s_server.
6899 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
6901 * X509_ATTRIBUTE memory leak
6903 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6904 memory. This structure is used by the PKCS#7 and CMS routines so any
6905 application which reads PKCS#7 or CMS data from untrusted sources is
6906 affected. SSL/TLS is not affected.
6908 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6914 * Race condition handling PSK identify hint
6916 If PSK identity hints are received by a multi-threaded client then
6917 the values are wrongly updated in the parent SSL_CTX structure. This can
6918 result in a race condition potentially leading to a double free of the
6924 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
6926 * Malformed ECParameters causes infinite loop
6928 When processing an ECParameters structure OpenSSL enters an infinite loop
6929 if the curve specified is over a specially malformed binary polynomial
6932 This can be used to perform denial of service against any
6933 system which processes public keys, certificate requests or
6934 certificates. This includes TLS clients and TLS servers with
6935 client authentication enabled.
6937 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6942 * Exploitable out-of-bounds read in X509_cmp_time
6944 X509_cmp_time does not properly check the length of the ASN1_TIME
6945 string and can read a few bytes out of bounds. In addition,
6946 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6949 An attacker can use this to craft malformed certificates and CRLs of
6950 various sizes and potentially cause a segmentation fault, resulting in
6951 a DoS on applications that verify certificates or CRLs. TLS clients
6952 that verify CRLs are affected. TLS clients and servers with client
6953 authentication enabled may be affected if they use custom verification
6956 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6957 independently by Hanno Böck.
6962 * PKCS7 crash with missing EnvelopedContent
6964 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6965 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6966 with missing content and trigger a NULL pointer dereference on parsing.
6968 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6969 structures from untrusted sources are affected. OpenSSL clients and
6970 servers are not affected.
6972 This issue was reported to OpenSSL by Michal Zalewski (Google).
6977 * CMS verify infinite loop with unknown hash function
6979 When verifying a signedData message the CMS code can enter an infinite loop
6980 if presented with an unknown hash function OID. This can be used to perform
6981 denial of service against any system which verifies signedData messages using
6983 This issue was reported to OpenSSL by Johannes Bauer.
6988 * Race condition handling NewSessionTicket
6990 If a NewSessionTicket is received by a multi-threaded client when attempting to
6991 reuse a previous ticket then a race condition can occur potentially leading to
6992 a double free of the ticket data.
6997 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
6999 * Segmentation fault in ASN1_TYPE_cmp fix
7001 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7002 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7003 certificate signature algorithm consistency this can be used to crash any
7004 certificate verification operation and exploited in a DoS attack. Any
7005 application which performs certificate verification is vulnerable including
7006 OpenSSL clients and servers which enable client authentication.
7011 * ASN.1 structure reuse memory corruption fix
7013 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7014 memory corruption via an invalid write. Such reuse is and has been
7015 strongly discouraged and is believed to be rare.
7017 Applications that parse structures containing CHOICE or ANY DEFINED BY
7018 components may be affected. Certificate parsing (d2i_X509 and related
7019 functions) are however not affected. OpenSSL clients and servers are
7025 * PKCS7 NULL pointer dereferences fix
7027 The PKCS#7 parsing code does not handle missing outer ContentInfo
7028 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7029 missing content and trigger a NULL pointer dereference on parsing.
7031 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7032 otherwise parse PKCS#7 structures from untrusted sources are
7033 affected. OpenSSL clients and servers are not affected.
7035 This issue was reported to OpenSSL by Michal Zalewski (Google).
7040 * DoS via reachable assert in SSLv2 servers fix
7042 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7043 servers that both support SSLv2 and enable export cipher suites by sending
7044 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7046 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7047 (OpenSSL development team).
7052 * Use After Free following d2i_ECPrivatekey error fix
7054 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7055 could cause a use after free condition. This, in turn, could cause a double
7056 free in several private key parsing functions (such as d2i_PrivateKey
7057 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7058 for applications that receive EC private keys from untrusted
7059 sources. This scenario is considered rare.
7061 This issue was discovered by the BoringSSL project and fixed in their
7067 * X509_to_X509_REQ NULL pointer deref fix
7069 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7070 the certificate key is invalid. This function is rarely used in practice.
7072 This issue was discovered by Brian Carpenter.
7077 * Removed the export ciphers from the DEFAULT ciphers
7081 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7083 * Build fixes for the Windows and OpenVMS platforms
7085 *Matt Caswell and Richard Levitte*
7087 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7089 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7090 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7091 dereference. This could lead to a Denial Of Service attack. Thanks to
7092 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7097 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7098 dtls1_buffer_record function under certain conditions. In particular this
7099 could occur if an attacker sent repeated DTLS records with the same
7100 sequence number but for the next epoch. The memory leak could be exploited
7101 by an attacker in a Denial of Service attack through memory exhaustion.
7102 Thanks to Chris Mueller for reporting this issue.
7107 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7108 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7109 method would be set to NULL which could later result in a NULL pointer
7110 dereference. Thanks to Frank Schmirler for reporting this issue.
7115 * Abort handshake if server key exchange message is omitted for ephemeral
7118 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7119 reporting this issue.
7124 * Remove non-export ephemeral RSA code on client and server. This code
7125 violated the TLS standard by allowing the use of temporary RSA keys in
7126 non-export ciphersuites and could be used by a server to effectively
7127 downgrade the RSA key length used to a value smaller than the server
7128 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7129 INRIA or reporting this issue.
7134 * Fixed issue where DH client certificates are accepted without verification.
7135 An OpenSSL server will accept a DH certificate for client authentication
7136 without the certificate verify message. This effectively allows a client to
7137 authenticate without the use of a private key. This only affects servers
7138 which trust a client certificate authority which issues certificates
7139 containing DH keys: these are extremely rare and hardly ever encountered.
7140 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7146 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7147 results on some platforms, including x86_64. This bug occurs at random
7148 with a very low probability, and is not known to be exploitable in any
7149 way, though its exact impact is difficult to determine. Thanks to Pieter
7150 Wuille (Blockstream) who reported this issue and also suggested an initial
7151 fix. Further analysis was conducted by the OpenSSL development team and
7152 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7153 the OpenSSL core team.
7158 * Fix various certificate fingerprint issues.
7160 By using non-DER or invalid encodings outside the signed portion of a
7161 certificate the fingerprint can be changed without breaking the signature.
7162 Although no details of the signed portion of the certificate can be changed
7163 this can cause problems with some applications: e.g. those using the
7164 certificate fingerprint for blacklists.
7166 1. Reject signatures with non zero unused bits.
7168 If the BIT STRING containing the signature has non zero unused bits reject
7169 the signature. All current signature algorithms require zero unused bits.
7171 2. Check certificate algorithm consistency.
7173 Check the AlgorithmIdentifier inside TBS matches the one in the
7174 certificate signature. NB: this will result in signature failure
7175 errors for some broken certificates.
7177 Thanks to Konrad Kraszewski from Google for reporting this issue.
7179 3. Check DSA/ECDSA signatures use DER.
7181 Reencode DSA/ECDSA signatures and compare with the original received
7182 signature. Return an error if there is a mismatch.
7184 This will reject various cases including garbage after signature
7185 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7186 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7187 (negative or with leading zeroes).
7189 Further analysis was conducted and fixes were developed by Stephen Henson
7190 of the OpenSSL core team.
7196 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7198 * Session Ticket Memory Leak.
7200 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7201 integrity of that ticket is first verified. In the event of a session
7202 ticket integrity check failing, OpenSSL will fail to free memory
7203 causing a memory leak. By sending a large number of invalid session
7204 tickets an attacker could exploit this issue in a Denial Of Service
7210 * Build option no-ssl3 is incomplete.
7212 When OpenSSL is configured with "no-ssl3" as a build option, servers
7213 could accept and complete a SSL 3.0 handshake, and clients could be
7214 configured to send them.
7217 *Akamai and the OpenSSL team*
7219 * Add support for TLS_FALLBACK_SCSV.
7220 Client applications doing fallback retries should call
7221 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7224 *Adam Langley, Bodo Moeller*
7226 * Add additional DigestInfo checks.
7228 Reencode DigestInto in DER and check against the original when
7229 verifying RSA signature: this will reject any improperly encoded
7230 DigestInfo structures.
7232 Note: this is a precautionary measure and no attacks are currently known.
7236 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7238 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7239 to a denial of service attack. A malicious server can crash the client
7240 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7241 ciphersuite and sending carefully crafted handshake messages.
7243 Thanks to Felix Gröbert (Google) for discovering and researching this
7249 * By sending carefully crafted DTLS packets an attacker could cause openssl
7250 to leak memory. This can be exploited through a Denial of Service attack.
7251 Thanks to Adam Langley for discovering and researching this issue.
7256 * An attacker can force openssl to consume large amounts of memory whilst
7257 processing DTLS handshake messages. This can be exploited through a
7258 Denial of Service attack.
7259 Thanks to Adam Langley for discovering and researching this issue.
7264 * An attacker can force an error condition which causes openssl to crash
7265 whilst processing DTLS packets due to memory being freed twice. This
7266 can be exploited through a Denial of Service attack.
7267 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7273 * If a multithreaded client connects to a malicious server using a resumed
7274 session and the server sends an ec point format extension it could write
7275 up to 255 bytes to freed memory.
7277 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7283 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7284 X509_name_oneline, X509_name_print_ex et al. to leak some information
7285 from the stack. Applications may be affected if they echo pretty printing
7286 output to the attacker.
7288 Thanks to Ivan Fratric (Google) for discovering this issue.
7291 *Emilia Käsper, and Steve Henson*
7293 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7294 for corner cases. (Certain input points at infinity could lead to
7295 bogus results, with non-infinity inputs mapped to infinity too.)
7299 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7301 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7302 handshake can force the use of weak keying material in OpenSSL
7303 SSL/TLS clients and servers.
7305 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7306 researching this issue. [CVE-2014-0224][]
7308 *KIKUCHI Masashi, Steve Henson*
7310 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7311 OpenSSL DTLS client the code can be made to recurse eventually crashing
7314 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7317 *Imre Rad, Steve Henson*
7319 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7320 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7321 client or server. This is potentially exploitable to run arbitrary
7322 code on a vulnerable client or server.
7324 Thanks to Jüri Aedla for reporting this issue. [CVE-2014-0195][]
7326 *Jüri Aedla, Steve Henson*
7328 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7329 are subject to a denial of service attack.
7331 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7332 this issue. [CVE-2014-3470][]
7334 *Felix Gröbert, Ivan Fratric, Steve Henson*
7336 * Harmonize version and its documentation. -f flag is used to display
7339 *mancha <mancha1@zoho.com>*
7341 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7342 in i2d_ECPrivateKey.
7344 *mancha <mancha1@zoho.com>*
7346 * Fix some double frees. These are not thought to be exploitable.
7348 *mancha <mancha1@zoho.com>*
7350 * Fix for the attack described in the paper "Recovering OpenSSL
7351 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7352 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7353 <http://eprint.iacr.org/2014/140>
7355 Thanks to Yuval Yarom and Naomi Benger for discovering this
7356 flaw and to Yuval Yarom for supplying a fix [CVE-2014-0076][]
7358 *Yuval Yarom and Naomi Benger*
7360 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7362 * Keep original DTLS digest and encryption contexts in retransmission
7363 structures so we can use the previous session parameters if they need
7364 to be resent. [CVE-2013-6450][]
7368 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7369 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7370 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7371 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7372 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7373 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7375 *Rob Stradling, Adam Langley*
7377 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7379 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7381 This addresses the flaw in CBC record processing discovered by
7382 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7383 at: <http://www.isg.rhul.ac.uk/tls/>
7385 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7386 Security Group at Royal Holloway, University of London
7387 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7388 Emilia Käsper for the initial patch.
7391 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7393 * Return an error when checking OCSP signatures when key is NULL.
7394 This fixes a DoS attack. [CVE-2013-0166][]
7398 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7399 the right response is stapled. Also change SSL_get_certificate()
7400 so it returns the certificate actually sent.
7401 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7402 (This is a backport)
7404 *Rob Stradling <rob.stradling@comodo.com>*
7406 * Fix possible deadlock when decoding public keys.
7410 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7412 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7415 * Sanity check record length before skipping explicit IV in DTLS
7418 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7419 fuzzing as a service testing platform.
7424 * Initialise tkeylen properly when encrypting CMS messages.
7425 Thanks to Solar Designer of Openwall for reporting this issue.
7429 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7431 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7432 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7433 in CRYPTO_realloc_clean.
7435 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7436 issue and to Adam Langley <agl@chromium.org> for fixing it.
7439 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7441 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7443 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7444 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7445 content decryption and always return the same error. Note: this attack
7446 needs on average 2^20 messages so it only affects automated senders. The
7447 old behaviour can be re-enabled in the CMS code by setting the
7448 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7449 an MMA defence is not necessary.
7450 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7451 this issue. [CVE-2012-0884][]
7455 * Fix CVE-2011-4619: make sure we really are receiving a
7456 client hello before rejecting multiple SGC restarts. Thanks to
7457 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7461 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7463 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7464 Thanks to Antonio Martin, Enterprise Secure Access Research and
7465 Development, Cisco Systems, Inc. for discovering this bug and
7466 preparing a fix. [CVE-2012-0050][]
7470 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7472 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7473 of the Vaudenay padding oracle attack on CBC mode encryption
7474 which enables an efficient plaintext recovery attack against
7475 the OpenSSL implementation of DTLS. Their attack exploits timing
7476 differences arising during decryption processing. A research
7477 paper describing this attack can be found at:
7478 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7479 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7480 Security Group at Royal Holloway, University of London
7481 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7482 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7483 for preparing the fix. [CVE-2011-4108][]
7485 *Robin Seggelmann, Michael Tuexen*
7487 * Clear bytes used for block padding of SSL 3.0 records.
7490 *Adam Langley (Google)*
7492 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7493 Kadianakis <desnacked@gmail.com> for discovering this issue and
7494 Adam Langley for preparing the fix. [CVE-2011-4619][]
7496 *Adam Langley (Google)*
7498 * Check parameters are not NULL in GOST ENGINE. [CVE-2012-0027][]
7500 *Andrey Kulikov <amdeich@gmail.com>*
7502 * Prevent malformed RFC3779 data triggering an assertion failure.
7503 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7504 and Rob Austein <sra@hactrn.net> for fixing it. [CVE-2011-4577][]
7506 *Rob Austein <sra@hactrn.net>*
7508 * Improved PRNG seeding for VOS.
7510 *Paul Green <Paul.Green@stratus.com>*
7512 * Fix ssl_ciph.c set-up race.
7514 *Adam Langley (Google)*
7516 * Fix spurious failures in ecdsatest.c.
7518 *Emilia Käsper (Google)*
7520 * Fix the BIO_f_buffer() implementation (which was mixing different
7521 interpretations of the `..._len` fields).
7523 *Adam Langley (Google)*
7525 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7526 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7527 threads won't reuse the same blinding coefficients.
7529 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7530 lock to call BN_BLINDING_invert_ex, and avoids one use of
7531 BN_BLINDING_update for each BN_BLINDING structure (previously,
7532 the last update always remained unused).
7534 *Emilia Käsper (Google)*
7536 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7538 *Bob Buckholz (Google)*
7540 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7542 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7543 by initialising X509_STORE_CTX properly. [CVE-2011-3207][]
7545 *Kaspar Brand <ossl@velox.ch>*
7547 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7548 for multi-threaded use of ECDH. [CVE-2011-3210][]
7550 *Adam Langley (Google)*
7552 * Fix x509_name_ex_d2i memory leak on bad inputs.
7556 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7557 signature public key algorithm by using OID xref utilities instead.
7558 Before this you could only use some ECC ciphersuites with SHA1 only.
7562 * Add protection against ECDSA timing attacks as mentioned in the paper
7563 by Billy Bob Brumley and Nicola Tuveri, see:
7564 <http://eprint.iacr.org/2011/232.pdf>
7566 *Billy Bob Brumley and Nicola Tuveri*
7568 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7570 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7572 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7574 * Fix bug in string printing code: if *any* escaping is enabled we must
7575 escape the escape character (backslash) or the resulting string is
7580 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7582 * Disable code workaround for ancient and obsolete Netscape browsers
7583 and servers: an attacker can use it in a ciphersuite downgrade attack.
7584 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7588 * Fixed J-PAKE implementation error, originally discovered by
7589 Sebastien Martini, further info and confirmation from Stefan
7590 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7594 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7596 * Fix extension code to avoid race conditions which can result in a buffer
7597 overrun vulnerability: resumed sessions must not be modified as they can
7598 be shared by multiple threads. CVE-2010-3864
7602 * Fix WIN32 build system to correctly link an ENGINE directory into
7607 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7609 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7612 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7614 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7616 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7617 context. The operation can be customised via the ctrl mechanism in
7618 case ENGINEs want to include additional functionality.
7622 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7626 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7627 output hashes compatible with older versions of OpenSSL.
7629 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7631 * Fix compression algorithm handling: if resuming a session use the
7632 compression algorithm of the resumed session instead of determining
7633 it from client hello again. Don't allow server to change algorithm.
7637 * Add load_crls() function to apps tidying load_certs() too. Add option
7638 to verify utility to allow additional CRLs to be included.
7642 * Update OCSP request code to permit adding custom headers to the request:
7643 some responders need this.
7647 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7650 *Julia Lawall <julia@diku.dk>*
7652 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7653 needlessly dereferenced structures, used obsolete functions and
7654 didn't handle all updated verify codes correctly.
7658 * Disable MD2 in the default configuration.
7662 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7663 indicate the initial BIO being pushed or popped. This makes it possible
7664 to determine whether the BIO is the one explicitly called or as a result
7665 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7666 it handles reference counts correctly and doesn't zero out the I/O bio
7667 when it is not being explicitly popped. WARNING: applications which
7668 included workarounds for the old buggy behaviour will need to be modified
7669 or they could free up already freed BIOs.
7673 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7674 renaming to all platforms (within the 0.9.8 branch, this was
7675 done conditionally on Netware platforms to avoid a name clash).
7677 *Guenter <lists@gknw.net>*
7679 * Add ECDHE and PSK support to DTLS.
7681 *Michael Tuexen <tuexen@fh-muenster.de>*
7683 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7688 * Add "missing" function EVP_MD_flags() (without this the only way to
7689 retrieve a digest flags is by accessing the structure directly. Update
7690 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7691 or cipher is registered as in the "from" argument. Print out all
7692 registered digests in the dgst usage message instead of manually
7693 attempting to work them out.
7697 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7698 this allows the use of compression and extensions. Change default cipher
7699 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7700 by default unless an application cipher string requests it.
7704 * Alter match criteria in PKCS12_parse(). It used to try to use local
7705 key ids to find matching certificates and keys but some PKCS#12 files
7706 don't follow the (somewhat unwritten) rules and this strategy fails.
7707 Now just gather all certificates together and the first private key
7708 then look for the first certificate that matches the key.
7712 * Support use of registered digest and cipher names for dgst and cipher
7713 commands instead of having to add each one as a special case. So now
7720 openssl dgst -sha256 foo
7722 and this works for ENGINE based algorithms too.
7726 * Update Gost ENGINE to support parameter files.
7728 *Victor B. Wagner <vitus@cryptocom.ru>*
7730 * Support GeneralizedTime in ca utility.
7732 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7734 * Enhance the hash format used for certificate directory links. The new
7735 form uses the canonical encoding (meaning equivalent names will work
7736 even if they aren't identical) and uses SHA1 instead of MD5. This form
7737 is incompatible with the older format and as a result c_rehash should
7738 be used to rebuild symbolic links.
7742 * Make PKCS#8 the default write format for private keys, replacing the
7743 traditional format. This form is standardised, more secure and doesn't
7744 include an implicit MD5 dependency.
7748 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7749 committed to OpenSSL should pass this lot as a minimum.
7753 * Add session ticket override functionality for use by EAP-FAST.
7755 *Jouni Malinen <j@w1.fi>*
7757 * Modify HMAC functions to return a value. Since these can be implemented
7758 in an ENGINE errors can occur.
7762 * Type-checked OBJ_bsearch_ex.
7766 * Type-checked OBJ_bsearch. Also some constification necessitated
7767 by type-checking. Still to come: TXT_DB, bsearch(?),
7768 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7773 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7774 seconds to a tm structure directly, instead of going through OS
7775 specific date routines. This avoids any issues with OS routines such
7776 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7777 and X509_time_adj_ex() to cover the extended range. The existing
7778 X509_time_adj() is still usable and will no longer have any date issues.
7782 * Delta CRL support. New use deltas option which will attempt to locate
7783 and search any appropriate delta CRLs available.
7785 This work was sponsored by Google.
7789 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7790 code and add additional score elements. Validate alternate CRL paths
7791 as part of the CRL checking and indicate a new error "CRL path validation
7792 error" in this case. Applications wanting additional details can use
7793 the verify callback and check the new "parent" field. If this is not
7794 NULL CRL path validation is taking place. Existing applications won't
7795 see this because it requires extended CRL support which is off by
7798 This work was sponsored by Google.
7802 * Support for freshest CRL extension.
7804 This work was sponsored by Google.
7808 * Initial indirect CRL support. Currently only supported in the CRLs
7809 passed directly and not via lookup. Process certificate issuer
7810 CRL entry extension and lookup CRL entries by bother issuer name
7811 and serial number. Check and process CRL issuer entry in IDP extension.
7813 This work was sponsored by Google.
7817 * Add support for distinct certificate and CRL paths. The CRL issuer
7818 certificate is validated separately in this case. Only enabled if
7819 an extended CRL support flag is set: this flag will enable additional
7820 CRL functionality in future.
7822 This work was sponsored by Google.
7826 * Add support for policy mappings extension.
7828 This work was sponsored by Google.
7832 * Fixes to pathlength constraint, self issued certificate handling,
7833 policy processing to align with RFC3280 and PKITS tests.
7835 This work was sponsored by Google.
7839 * Support for name constraints certificate extension. DN, email, DNS
7840 and URI types are currently supported.
7842 This work was sponsored by Google.
7846 * To cater for systems that provide a pointer-based thread ID rather
7847 than numeric, deprecate the current numeric thread ID mechanism and
7848 replace it with a structure and associated callback type. This
7849 mechanism allows a numeric "hash" to be extracted from a thread ID in
7850 either case, and on platforms where pointers are larger than 'long',
7851 mixing is done to help ensure the numeric 'hash' is usable even if it
7852 can't be guaranteed unique. The default mechanism is to use "&errno"
7853 as a pointer-based thread ID to distinguish between threads.
7855 Applications that want to provide their own thread IDs should now use
7856 CRYPTO_THREADID_set_callback() to register a callback that will call
7857 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
7859 Note that ERR_remove_state() is now deprecated, because it is tied
7860 to the assumption that thread IDs are numeric. ERR_remove_state(0)
7861 to free the current thread's error state should be replaced by
7862 ERR_remove_thread_state(NULL).
7864 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
7865 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
7866 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
7867 application was previously providing a numeric thread callback that
7868 was inappropriate for distinguishing threads, then uniqueness might
7869 have been obtained with &errno that happened immediately in the
7870 intermediate development versions of OpenSSL; this is no longer the
7871 case, the numeric thread callback will now override the automatic use
7874 *Geoff Thorpe, with help from Bodo Moeller*
7876 * Initial support for different CRL issuing certificates. This covers a
7877 simple case where the self issued certificates in the chain exist and
7878 the real CRL issuer is higher in the existing chain.
7880 This work was sponsored by Google.
7884 * Removed effectively defunct crypto/store from the build.
7888 * Revamp of STACK to provide stronger type-checking. Still to come:
7889 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
7890 ASN1_STRING, CONF_VALUE.
7894 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
7895 RAM on SSL connections. This option can save about 34k per idle SSL.
7899 * Revamp of LHASH to provide stronger type-checking. Still to come:
7900 STACK, TXT_DB, bsearch, qsort.
7904 * Initial support for Cryptographic Message Syntax (aka CMS) based
7905 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
7906 support for data, signedData, compressedData, digestedData and
7907 encryptedData, envelopedData types included. Scripts to check against
7908 RFC4134 examples draft and interop and consistency checks of many
7909 content types and variants.
7913 * Add options to enc utility to support use of zlib compression BIO.
7917 * Extend mk1mf to support importing of options and assembly language
7918 files from Configure script, currently only included in VC-WIN32.
7919 The assembly language rules can now optionally generate the source
7920 files from the associated perl scripts.
7924 * Implement remaining functionality needed to support GOST ciphersuites.
7925 Interop testing has been performed using CryptoPro implementations.
7927 *Victor B. Wagner <vitus@cryptocom.ru>*
7929 * s390x assembler pack.
7933 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
7938 * Implement Opaque PRF Input TLS extension as specified in
7939 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
7940 official specification yet and no extension type assignment by
7941 IANA exists, this extension (for now) will have to be explicitly
7942 enabled when building OpenSSL by providing the extension number
7943 to use. For example, specify an option
7945 -DTLSEXT_TYPE_opaque_prf_input=0x9527
7947 to the "config" or "Configure" script to enable the extension,
7948 assuming extension number 0x9527 (which is a completely arbitrary
7949 and unofficial assignment based on the MD5 hash of the Internet
7950 Draft). Note that by doing so, you potentially lose
7951 interoperability with other TLS implementations since these might
7952 be using the same extension number for other purposes.
7954 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
7955 opaque PRF input value to use in the handshake. This will create
7956 an internal copy of the length-'len' string at 'src', and will
7957 return non-zero for success.
7959 To get more control and flexibility, provide a callback function
7962 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
7963 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
7967 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
7970 Callback function 'cb' will be called in handshakes, and is
7971 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
7972 Argument 'arg' is for application purposes (the value as given to
7973 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
7974 be provided to the callback function). The callback function
7975 has to return non-zero to report success: usually 1 to use opaque
7976 PRF input just if possible, or 2 to enforce use of the opaque PRF
7977 input. In the latter case, the library will abort the handshake
7978 if opaque PRF input is not successfully negotiated.
7980 Arguments 'peerinput' and 'len' given to the callback function
7981 will always be NULL and 0 in the case of a client. A server will
7982 see the client's opaque PRF input through these variables if
7983 available (NULL and 0 otherwise). Note that if the server
7984 provides an opaque PRF input, the length must be the same as the
7985 length of the client's opaque PRF input.
7987 Note that the callback function will only be called when creating
7988 a new session (session resumption can resume whatever was
7989 previously negotiated), and will not be called in SSL 2.0
7990 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
7991 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
7992 for applications that need to enforce opaque PRF input.
7996 * Update ssl code to support digests other than SHA1+MD5 for handshake
7999 *Victor B. Wagner <vitus@cryptocom.ru>*
8001 * Add RFC4507 support to OpenSSL. This includes the corrections in
8002 RFC4507bis. The encrypted ticket format is an encrypted encoded
8003 SSL_SESSION structure, that way new session features are automatically
8006 If a client application caches session in an SSL_SESSION structure
8007 support is transparent because tickets are now stored in the encoded
8010 The SSL_CTX structure automatically generates keys for ticket
8011 protection in servers so again support should be possible
8012 with no application modification.
8014 If a client or server wishes to disable RFC4507 support then the option
8015 SSL_OP_NO_TICKET can be set.
8017 Add a TLS extension debugging callback to allow the contents of any client
8018 or server extensions to be examined.
8020 This work was sponsored by Google.
8024 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8025 OpenSSL should now compile cleanly on gcc 4.2
8027 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8029 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8030 support including streaming MAC support: this is required for GOST
8031 ciphersuite support.
8033 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8035 * Add option -stream to use PKCS#7 streaming in smime utility. New
8036 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8037 to output in BER and PEM format.
8041 * Experimental support for use of HMAC via EVP_PKEY interface. This
8042 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8043 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8044 ENGINE support for HMAC keys which are unextractable. New -mac and
8045 -macopt options to dgst utility.
8049 * New option -sigopt to dgst utility. Update dgst to use
8050 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8051 alternative signing parameters such as X9.31 or PSS in the dgst
8056 * Change ssl_cipher_apply_rule(), the internal function that does
8057 the work each time a ciphersuite string requests enabling
8058 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8059 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8060 the order of disabled ciphersuites such that those ciphersuites
8061 that most recently went from enabled to disabled not only stay
8062 in order with respect to each other, but also have higher priority
8063 than other disabled ciphersuites the next time ciphersuites are
8066 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8067 the same ciphersuites as with "HIGH" alone, but in a specific
8068 order where the PSK ciphersuites come first (since they are the
8069 most recently disabled ciphersuites when "HIGH" is parsed).
8071 Also, change ssl_create_cipher_list() (using this new
8072 functionality) such that between otherwise identical
8073 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8078 * Change ssl_create_cipher_list() so that it automatically
8079 arranges the ciphersuites in reasonable order before starting
8080 to process the rule string. Thus, the definition for "DEFAULT"
8081 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8082 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8083 This makes it much easier to arrive at a reasonable default order
8084 in applications for which anonymous ciphers are OK (meaning
8085 that you can't actually use DEFAULT).
8087 *Bodo Moeller; suggested by Victor Duchovni*
8089 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8090 processing) into multiple integers instead of setting
8091 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8092 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8093 (These masks as well as the individual bit definitions are hidden
8094 away into the non-exported interface ssl/ssl_locl.h, so this
8095 change to the definition of the SSL_CIPHER structure shouldn't
8096 affect applications.) This give us more bits for each of these
8097 categories, so there is no longer a need to coagulate AES128 and
8098 AES256 into a single algorithm bit, and to coagulate Camellia128
8099 and Camellia256 into a single algorithm bit, which has led to all
8102 Thus, among other things, the kludge introduced in 0.9.7m and
8103 0.9.8e for masking out AES256 independently of AES128 or masking
8104 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8106 With the change, we also introduce new ciphersuite aliases that
8107 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8112 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8113 Use the leftmost N bytes of the signature input if the input is
8114 larger than the prime q (with N being the size in bytes of q).
8118 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8119 it yet and it is largely untested.
8123 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8127 * Initial incomplete changes to avoid need for function casts in OpenSSL
8128 some compilers (gcc 4.2 and later) reject their use. Safestack is
8129 reimplemented. Update ASN1 to avoid use of legacy functions.
8133 * Win32/64 targets are linked with Winsock2.
8137 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8138 to external functions. This can be used to increase CRL handling
8139 efficiency especially when CRLs are very large by (for example) storing
8140 the CRL revoked certificates in a database.
8144 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8145 new CRLs added to a directory can be used. New command line option
8146 -verify_return_error to s_client and s_server. This causes real errors
8147 to be returned by the verify callback instead of carrying on no matter
8148 what. This reflects the way a "real world" verify callback would behave.
8152 * GOST engine, supporting several GOST algorithms and public key formats.
8153 Kindly donated by Cryptocom.
8157 * Partial support for Issuing Distribution Point CRL extension. CRLs
8158 partitioned by DP are handled but no indirect CRL or reason partitioning
8159 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8160 selected via a scoring technique which handles IDP and AKID in CRLs.
8164 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8165 will ultimately be used for all verify operations: this will remove the
8166 X509_STORE dependency on certificate verification and allow alternative
8167 lookup methods. X509_STORE based implementations of these two callbacks.
8171 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8172 Modify get_crl() to find a valid (unexpired) CRL if possible.
8176 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8177 this would be called X509_CRL_cmp() but that name is already used by
8178 a function that just compares CRL issuer names. Cache several CRL
8179 extensions in X509_CRL structure and cache CRLDP in X509.
8183 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8184 this maps equivalent X509_NAME structures into a consistent structure.
8185 Name comparison can then be performed rapidly using memcmp().
8189 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8194 * Allow digests to supply their own micalg string for S/MIME type using
8195 the ctrl EVP_MD_CTRL_MICALG.
8199 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8200 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8201 ctrl. It can then customise the structure before and/or after signing
8206 * New function OBJ_add_sigid() to allow application defined signature OIDs
8207 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8208 to free up any added signature OIDs.
8212 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8213 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8214 digest and cipher tables. New options added to openssl utility:
8215 list-message-digest-algorithms and list-cipher-algorithms.
8219 * Change the array representation of binary polynomials: the list
8220 of degrees of non-zero coefficients is now terminated with -1.
8221 Previously it was terminated with 0, which was also part of the
8222 value; thus, the array representation was not applicable to
8223 polynomials where t^0 has coefficient zero. This change makes
8224 the array representation useful in a more general context.
8228 * Various modifications and fixes to SSL/TLS cipher string
8229 handling. For ECC, the code now distinguishes between fixed ECDH
8230 with RSA certificates on the one hand and with ECDSA certificates
8231 on the other hand, since these are separate ciphersuites. The
8232 unused code for Fortezza ciphersuites has been removed.
8234 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8235 (not "ECDHE"). For consistency with the code for DH
8236 certificates, use of ECDH certificates is now considered ECDH
8237 authentication, not RSA or ECDSA authentication (the latter is
8238 merely the CA's signing algorithm and not actively used in the
8241 The temporary ciphersuite alias "ECCdraft" is no longer
8242 available, and ECC ciphersuites are no longer excluded from "ALL"
8243 and "DEFAULT". The following aliases now exist for RFC 4492
8244 ciphersuites, most of these by analogy with the DH case:
8246 kECDHr - ECDH cert, signed with RSA
8247 kECDHe - ECDH cert, signed with ECDSA
8248 kECDH - ECDH cert (signed with either RSA or ECDSA)
8249 kEECDH - ephemeral ECDH
8250 ECDH - ECDH cert or ephemeral ECDH
8256 AECDH - anonymous ECDH
8257 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8261 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8262 Use correct micalg parameters depending on digest(s) in signed message.
8266 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8267 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8271 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8272 an engine to register a method. Add ENGINE lookups for methods and
8273 functional reference processing.
8277 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8278 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8283 * New -resign option to smime utility. This adds one or more signers
8284 to an existing PKCS#7 signedData structure. Also -md option to use an
8285 alternative message digest algorithm for signing.
8289 * Tidy up PKCS#7 routines and add new functions to make it easier to
8290 create PKCS7 structures containing multiple signers. Update smime
8291 application to support multiple signers.
8295 * New -macalg option to pkcs12 utility to allow setting of an alternative
8300 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8301 Reorganize PBE internals to lookup from a static table using NIDs,
8302 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8303 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8304 PRF which will be automatically used with PBES2.
8308 * Replace the algorithm specific calls to generate keys in "req" with the
8313 * Update PKCS#7 enveloped data routines to use new API. This is now
8314 supported by any public key method supporting the encrypt operation. A
8315 ctrl is added to allow the public key algorithm to examine or modify
8316 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8321 * Add a ctrl to asn1 method to allow a public key algorithm to express
8322 a default digest type to use. In most cases this will be SHA1 but some
8323 algorithms (such as GOST) need to specify an alternative digest. The
8324 return value indicates how strong the preference is 1 means optional and
8325 2 is mandatory (that is it is the only supported type). Modify
8326 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8327 use the default md. Update openssl utilities to use the default digest
8328 type for signing if it is not explicitly indicated.
8332 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8333 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8334 signing method from the key type. This effectively removes the link
8335 between digests and public key types.
8339 * Add an OID cross reference table and utility functions. Its purpose is to
8340 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8341 rsaEncryption. This will allow some of the algorithm specific hackery
8342 needed to use the correct OID to be removed.
8346 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8347 structures for PKCS7_sign(). They are now set up by the relevant public
8352 * Add provisional EC pkey method with support for ECDSA and ECDH.
8356 * Add support for key derivation (agreement) in the API, DH method and
8361 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8362 public and private key formats. As a side effect these add additional
8363 command line functionality not previously available: DSA signatures can be
8364 generated and verified using pkeyutl and DH key support and generation in
8371 *Oliver Tappe <zooey@hirschkaefer.de>*
8373 * New make target "install_html_docs" installs HTML renditions of the
8376 *Oliver Tappe <zooey@hirschkaefer.de>*
8378 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8379 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8380 support key and parameter generation and add initial key generation
8381 functionality for RSA.
8385 * Add functions for main EVP_PKEY_method operations. The undocumented
8386 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8387 `EVP_PKEY_{encrypt,decrypt}_old`.
8391 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8392 key API, doesn't do much yet.
8396 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8397 public key algorithms. New option to openssl utility:
8398 "list-public-key-algorithms" to print out info.
8402 * Implement the Supported Elliptic Curves Extension for
8403 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8407 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8408 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8412 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8413 utilities such as rsa, dsa, dsaparam etc except they process any key
8418 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8419 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8420 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8425 * Initial support for pluggable public key ASN1.
8426 De-spaghettify the public key ASN1 handling. Move public and private
8427 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8428 algorithm specific handling to a single module within the relevant
8429 algorithm directory. Add functions to allow (near) opaque processing
8430 of public and private key structures.
8434 * Implement the Supported Point Formats Extension for
8435 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8439 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8440 for the psk identity [hint] and the psk callback functions to the
8441 SSL_SESSION, SSL and SSL_CTX structure.
8444 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8448 SSL_CTX_use_psk_identity_hint
8449 SSL_get_psk_identity_hint
8450 SSL_get_psk_identity
8451 SSL_use_psk_identity_hint
8453 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8455 * Add RFC 3161 compliant time stamp request creation, response generation
8456 and response verification functionality.
8458 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8460 * Add initial support for TLS extensions, specifically for the server_name
8461 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8462 have new members for a host name. The SSL data structure has an
8463 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8464 stored in that context to allow for session resumption, even after the
8465 SSL has been switched to a new SSL_CTX in reaction to a client's
8466 server_name extension.
8468 New functions (subject to change):
8470 SSL_get_servername()
8471 SSL_get_servername_type()
8474 New CTRL codes and macros (subject to change):
8476 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8477 - SSL_CTX_set_tlsext_servername_callback()
8478 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8479 - SSL_CTX_set_tlsext_servername_arg()
8480 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8482 openssl s_client has a new '-servername ...' option.
8484 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8485 '-key2 ...', '-servername_fatal' (subject to change). This allows
8486 testing the HostName extension for a specific single host name ('-cert'
8487 and '-key' remain fallbacks for handshakes without HostName
8488 negotiation). If the unrecognized_name alert has to be sent, this by
8489 default is a warning; it becomes fatal with the '-servername_fatal'
8492 *Peter Sylvester, Remy Allais, Christophe Renou*
8494 * Whirlpool hash implementation is added.
8498 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8499 bn(64,32). Because of instruction set limitations it doesn't have
8500 any negative impact on performance. This was done mostly in order
8501 to make it possible to share assembler modules, such as bn_mul_mont
8502 implementations, between 32- and 64-bit builds without hassle.
8506 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8507 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8512 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8513 dedicated Montgomery multiplication procedure, is introduced.
8514 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8515 "64-bit" performance on certain 32-bit targets.
8519 * New option SSL_OP_NO_COMP to disable use of compression selectively
8520 in SSL structures. New SSL ctrl to set maximum send fragment size.
8521 Save memory by setting the I/O buffer sizes dynamically instead of
8522 using the maximum available value.
8526 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8527 in addition to the text details.
8531 * Very, very preliminary EXPERIMENTAL support for printing of general
8532 ASN1 structures. This currently produces rather ugly output and doesn't
8533 handle several customised structures at all.
8537 * Integrated support for PVK file format and some related formats such
8538 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8539 these in the 'rsa' and 'dsa' utilities.
8543 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8547 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8548 place for the (very old) "NETSCAPE" format certificates which are now
8549 handled using new ASN1 code equivalents.
8553 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8554 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8555 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8559 * Modify CRL distribution points extension code to print out previously
8560 unsupported fields. Enhance extension setting code to allow setting of
8565 * Add print and set support for Issuing Distribution Point CRL extension.
8569 * Change 'Configure' script to enable Camellia by default.
8576 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8578 * When rejecting SSL/TLS records due to an incorrect version number, never
8579 update s->server with a new major version number. As of
8580 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8581 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8582 the previous behavior could result in a read attempt at NULL when
8583 receiving specific incorrect SSL/TLS records once record payload
8584 protection is active. [CVE-2010-0740][]
8586 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8588 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8589 could be crashed if the relevant tables were not present (e.g. chrooted).
8591 *Tomas Hoger <thoger@redhat.com>*
8593 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8595 * Always check bn_wexpand() return values for failure. [CVE-2009-3245][]
8597 *Martin Olsson, Neel Mehta*
8599 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8600 accommodate for stack sorting, always a write lock!).
8604 * On some versions of WIN32 Heap32Next is very slow. This can cause
8605 excessive delays in the RAND_poll(): over a minute. As a workaround
8606 include a time check in the inner Heap32Next loop too.
8610 * The code that handled flushing of data in SSL/TLS originally used the
8611 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8612 the problem outlined in PR#1949. The fix suggested there however can
8613 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8614 of Apache). So instead simplify the code to flush unconditionally.
8615 This should be fine since flushing with no data to flush is a no op.
8619 * Handle TLS versions 2.0 and later properly and correctly use the
8620 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8621 off ancient servers have a habit of sticking around for a while...
8625 * Modify compression code so it frees up structures without using the
8626 ex_data callbacks. This works around a problem where some applications
8627 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8628 restarting) then use compression (e.g. SSL with compression) later.
8629 This results in significant per-connection memory leaks and
8630 has caused some security issues including CVE-2008-1678 and
8635 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8636 change when encrypting or decrypting.
8640 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8641 connect and renegotiate with servers which do not support RI.
8642 Until RI is more widely deployed this option is enabled by default.
8646 * Add "missing" ssl ctrls to clear options and mode.
8650 * If client attempts to renegotiate and doesn't support RI respond with
8651 a no_renegotiation alert as required by RFC5746. Some renegotiating
8652 TLS clients will continue a connection gracefully when they receive
8653 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8654 waiting for a server hello which it will never receive. Now we treat a
8655 received no_renegotiation alert as a fatal error. This is because
8656 applications requesting a renegotiation might well expect it to succeed
8657 and would have no code in place to handle the server denying it so the
8658 only safe thing to do is to terminate the connection.
8662 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8663 peer supports secure renegotiation and 0 otherwise. Print out peer
8664 renegotiation support in s_client/s_server.
8668 * Replace the highly broken and deprecated SPKAC certification method with
8669 the updated NID creation version. This should correctly handle UTF8.
8673 * Implement RFC5746. Re-enable renegotiation but require the extension
8674 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8675 turns out to be a bad idea. It has been replaced by
8676 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8677 SSL_CTX_set_options(). This is really not recommended unless you
8678 know what you are doing.
8680 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8682 * Fixes to stateless session resumption handling. Use initial_ctx when
8683 issuing and attempting to decrypt tickets in case it has changed during
8684 servername handling. Use a non-zero length session ID when attempting
8685 stateless session resumption: this makes it possible to determine if
8686 a resumption has occurred immediately after receiving server hello
8687 (several places in OpenSSL subtly assume this) instead of later in
8692 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8693 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8694 fixes for a few places where the return code is not checked
8697 *Julia Lawall <julia@diku.dk>*
8699 * Add --strict-warnings option to Configure script to include devteam
8700 warnings in other configurations.
8704 * Add support for --libdir option and LIBDIR variable in makefiles. This
8705 makes it possible to install openssl libraries in locations which
8706 have names other than "lib", for example "/usr/lib64" which some
8709 *Steve Henson, based on patch from Jeremy Utley*
8711 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8712 X690 8.9.12 and can produce some misleading textual output of OIDs.
8714 *Steve Henson, reported by Dan Kaminsky*
8716 * Delete MD2 from algorithm tables. This follows the recommendation in
8717 several standards that it is not used in new applications due to
8718 several cryptographic weaknesses. For binary compatibility reasons
8719 the MD2 API is still compiled in by default.
8723 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8728 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8729 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8732 *Guenter <lists@gknw.net>*
8734 * Fix the server certificate chain building code to use X509_verify_cert(),
8735 it used to have an ad-hoc builder which was unable to cope with anything
8736 other than a simple chain.
8738 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8740 * Don't check self signed certificate signatures in X509_verify_cert()
8741 by default (a flag can override this): it just wastes time without
8742 adding any security. As a useful side effect self signed root CAs
8743 with non-FIPS digests are now usable in FIPS mode.
8747 * In dtls1_process_out_of_seq_message() the check if the current message
8748 is already buffered was missing. For every new message was memory
8749 allocated, allowing an attacker to perform an denial of service attack
8750 with sending out of seq handshake messages until there is no memory
8751 left. Additionally every future message was buffered, even if the
8752 sequence number made no sense and would be part of another handshake.
8753 So only messages with sequence numbers less than 10 in advance will be
8754 buffered. [CVE-2009-1378][]
8756 *Robin Seggelmann, discovered by Daniel Mentz*
8758 * Records are buffered if they arrive with a future epoch to be
8759 processed after finishing the corresponding handshake. There is
8760 currently no limitation to this buffer allowing an attacker to perform
8761 a DOS attack with sending records with future epochs until there is no
8762 memory left. This patch adds the pqueue_size() function to determine
8763 the size of a buffer and limits the record buffer to 100 entries.
8766 *Robin Seggelmann, discovered by Daniel Mentz*
8768 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8769 parent structure is freed. [CVE-2009-1379][]
8773 * Handle non-blocking I/O properly in SSL_shutdown() call.
8775 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8777 * Add `2.5.4.*` OIDs
8779 *Ilya O. <vrghost@gmail.com>*
8781 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8783 * Disable renegotiation completely - this fixes a severe security
8784 problem [CVE-2009-3555][] at the cost of breaking all
8785 renegotiation. Renegotiation can be re-enabled by setting
8786 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8787 run-time. This is really not recommended unless you know what
8792 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8794 * Don't set val to NULL when freeing up structures, it is freed up by
8795 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
8796 zeroing past the valid field. [CVE-2009-0789][]
8798 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8800 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8801 checked correctly. This would allow some invalid signed attributes to
8802 appear to verify correctly. [CVE-2009-0591][]
8804 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8806 * Reject UniversalString and BMPString types with invalid lengths. This
8807 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8808 a legal length. [CVE-2009-0590][]
8812 * Set S/MIME signing as the default purpose rather than setting it
8813 unconditionally. This allows applications to override it at the store
8818 * Permit restricted recursion of ASN1 strings. This is needed in practice
8819 to handle some structures.
8823 * Improve efficiency of mem_gets: don't search whole buffer each time
8826 *Jeremy Shapiro <jnshapir@us.ibm.com>*
8828 * New -hex option for openssl rand.
8832 * Print out UTF8String and NumericString when parsing ASN1.
8836 * Support NumericString type for name components.
8840 * Allow CC in the environment to override the automatically chosen
8841 compiler. Note that nothing is done to ensure flags work with the
8846 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
8848 * Properly check EVP_VerifyFinal() and similar return values
8851 *Ben Laurie, Bodo Moeller, Google Security Team*
8853 * Enable TLS extensions by default.
8857 * Allow the CHIL engine to be loaded, whether the application is
8858 multithreaded or not. (This does not release the developer from the
8859 obligation to set up the dynamic locking callbacks.)
8861 *Sander Temme <sander@temme.net>*
8863 * Use correct exit code if there is an error in dgst command.
8865 *Steve Henson; problem pointed out by Roland Dirlewanger*
8867 * Tweak Configure so that you need to say "experimental-jpake" to enable
8868 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
8872 * Add experimental JPAKE support, including demo authentication in
8873 s_client and s_server.
8877 * Set the comparison function in v3_addr_canonize().
8879 *Rob Austein <sra@hactrn.net>*
8881 * Add support for XMPP STARTTLS in s_client.
8883 *Philip Paeps <philip@freebsd.org>*
8885 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
8886 to ensure that even with this option, only ciphersuites in the
8887 server's preference list will be accepted. (Note that the option
8888 applies only when resuming a session, so the earlier behavior was
8889 just about the algorithm choice for symmetric cryptography.)
8893 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
8895 * Fix NULL pointer dereference if a DTLS server received
8896 ChangeCipherSpec as first record [CVE-2009-1386][].
8900 * Fix a state transition in s3_srvr.c and d1_srvr.c
8901 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
8905 * The fix in 0.9.8c that supposedly got rid of unsafe
8906 double-checked locking was incomplete for RSA blinding,
8907 addressing just one layer of what turns out to have been
8908 doubly unsafe triple-checked locking.
8910 So now fix this for real by retiring the MONT_HELPER macro
8911 in crypto/rsa/rsa_eay.c.
8913 *Bodo Moeller; problem pointed out by Marius Schilder*
8915 * Various precautionary measures:
8917 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
8919 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
8920 (NB: This would require knowledge of the secret session ticket key
8921 to exploit, in which case you'd be SOL either way.)
8923 - Change bn_nist.c so that it will properly handle input BIGNUMs
8924 outside the expected range.
8926 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
8929 *Neel Mehta, Bodo Moeller*
8931 * Allow engines to be "soft loaded" - i.e. optionally don't die if
8932 the load fails. Useful for distros.
8934 *Ben Laurie and the FreeBSD team*
8936 * Add support for Local Machine Keyset attribute in PKCS#12 files.
8940 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
8944 * Expand ENGINE to support engine supplied SSL client certificate functions.
8946 This work was sponsored by Logica.
8950 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
8951 keystores. Support for SSL/TLS client authentication too.
8952 Not compiled unless enable-capieng specified to Configure.
8954 This work was sponsored by Logica.
8958 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
8959 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
8960 attribute creation routines such as certificate requests and PKCS#12
8965 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
8967 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
8968 handshake which could lead to a client crash as found using the
8969 Codenomicon TLS test suite [CVE-2008-1672][]
8971 *Steve Henson, Mark Cox*
8973 * Fix double free in TLS server name extensions which could lead to
8974 a remote crash found by Codenomicon TLS test suite [CVE-2008-0891][]
8978 * Clear error queue in SSL_CTX_use_certificate_chain_file()
8980 Clear the error queue to ensure that error entries left from
8981 older function calls do not interfere with the correct operation.
8983 *Lutz Jaenicke, Erik de Castro Lopo*
8985 * Remove root CA certificates of commercial CAs:
8987 The OpenSSL project does not recommend any specific CA and does not
8988 have any policy with respect to including or excluding any CA.
8989 Therefore it does not make any sense to ship an arbitrary selection
8990 of root CA certificates with the OpenSSL software.
8994 * RSA OAEP patches to fix two separate invalid memory reads.
8995 The first one involves inputs when 'lzero' is greater than
8996 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
8997 before the beginning of from). The second one involves inputs where
8998 the 'db' section contains nothing but zeroes (there is a one-byte
8999 invalid read after the end of 'db').
9001 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9003 * Partial backport from 0.9.9-dev:
9005 Introduce bn_mul_mont (dedicated Montgomery multiplication
9006 procedure) as a candidate for BIGNUM assembler implementation.
9007 While 0.9.9-dev uses assembler for various architectures, only
9008 x86_64 is available by default here in the 0.9.8 branch, and
9009 32-bit x86 is available through a compile-time setting.
9011 To try the 32-bit x86 assembler implementation, use Configure
9012 option "enable-montasm" (which exists only for this backport).
9014 As "enable-montasm" for 32-bit x86 disclaims code stability
9015 anyway, in this constellation we activate additional code
9016 backported from 0.9.9-dev for further performance improvements,
9017 namely BN_from_montgomery_word. (To enable this otherwise,
9018 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9020 *Andy Polyakov (backport partially by Bodo Moeller)*
9022 * Add TLS session ticket callback. This allows an application to set
9023 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9024 values. This is useful for key rollover for example where several key
9025 sets may exist with different names.
9029 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9030 This was broken until now in 0.9.8 releases, such that the only way
9031 a registered ENGINE could be used (assuming it initialises
9032 successfully on the host) was to explicitly set it as the default
9033 for the relevant algorithms. This is in contradiction with 0.9.7
9034 behaviour and the documentation. With this fix, when an ENGINE is
9035 registered into a given algorithm's table of implementations, the
9036 'uptodate' flag is reset so that auto-discovery will be used next
9037 time a new context for that algorithm attempts to select an
9040 *Ian Lister (tweaked by Geoff Thorpe)*
9042 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9043 implementation in the following ways:
9045 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9048 Lack of BER streaming support means one pass streaming processing is
9049 only supported if data is detached: setting the streaming flag is
9050 ignored for embedded content.
9052 CMS support is disabled by default and must be explicitly enabled
9053 with the enable-cms configuration option.
9057 * Update the GMP engine glue to do direct copies between BIGNUM and
9058 mpz_t when openssl and GMP use the same limb size. Otherwise the
9059 existing "conversion via a text string export" trick is still used.
9061 *Paul Sheer <paulsheer@gmail.com>*
9063 * Zlib compression BIO. This is a filter BIO which compressed and
9064 uncompresses any data passed through it.
9068 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9069 RFC3394 compatible AES key wrapping.
9073 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9074 sets string data without copying. X509_ALGOR_set0() and
9075 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9076 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9077 from an X509_ATTRIBUTE structure optionally checking it occurs only
9078 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9083 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9084 to get the expected BN_FLG_CONSTTIME behavior.
9086 *Bodo Moeller (Google)*
9090 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9091 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9092 - added some more tests to do_tests.pl
9093 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9094 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9095 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9096 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9097 - various changes to netware.pl to enable gcc-cross builds on Win32
9099 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9100 - various changes to fix missing prototype warnings
9101 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9102 - added AES, WHIRLPOOL and CPUID assembler code to build files
9103 - added missing AES assembler make rules to mk1mf.pl
9104 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9106 *Guenter Knauf <eflash@gmx.net>*
9108 * Implement certificate status request TLS extension defined in RFC3546.
9109 A client can set the appropriate parameters and receive the encoded
9110 OCSP response via a callback. A server can query the supplied parameters
9111 and set the encoded OCSP response in the callback. Add simplified examples
9112 to s_client and s_server.
9116 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9119 + Binary incompatibility of ssl_ctx_st structure
9120 + DTLS interoperation with non-compliant servers
9121 + Don't call get_session_cb() without proposed session
9122 + Fix ia64 assembler code
9124 *Andy Polyakov, Steve Henson*
9126 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9128 * DTLS Handshake overhaul. There were longstanding issues with
9129 OpenSSL DTLS implementation, which were making it impossible for
9130 RFC 4347 compliant client to communicate with OpenSSL server.
9131 Unfortunately just fixing these incompatibilities would "cut off"
9132 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9133 server keeps tolerating non RFC compliant syntax. The opposite is
9134 not true, 0.9.8f client can not communicate with earlier server.
9135 This update even addresses CVE-2007-4995.
9139 * Changes to avoid need for function casts in OpenSSL: some compilers
9140 (gcc 4.2 and later) reject their use.
9141 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9144 * Add RFC4507 support to OpenSSL. This includes the corrections in
9145 RFC4507bis. The encrypted ticket format is an encrypted encoded
9146 SSL_SESSION structure, that way new session features are automatically
9149 If a client application caches session in an SSL_SESSION structure
9150 support is transparent because tickets are now stored in the encoded
9153 The SSL_CTX structure automatically generates keys for ticket
9154 protection in servers so again support should be possible
9155 with no application modification.
9157 If a client or server wishes to disable RFC4507 support then the option
9158 SSL_OP_NO_TICKET can be set.
9160 Add a TLS extension debugging callback to allow the contents of any client
9161 or server extensions to be examined.
9163 This work was sponsored by Google.
9167 * Add initial support for TLS extensions, specifically for the server_name
9168 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9169 have new members for a host name. The SSL data structure has an
9170 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9171 stored in that context to allow for session resumption, even after the
9172 SSL has been switched to a new SSL_CTX in reaction to a client's
9173 server_name extension.
9175 New functions (subject to change):
9177 SSL_get_servername()
9178 SSL_get_servername_type()
9181 New CTRL codes and macros (subject to change):
9183 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9184 - SSL_CTX_set_tlsext_servername_callback()
9185 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9186 - SSL_CTX_set_tlsext_servername_arg()
9187 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9189 openssl s_client has a new '-servername ...' option.
9191 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9192 '-key2 ...', '-servername_fatal' (subject to change). This allows
9193 testing the HostName extension for a specific single host name ('-cert'
9194 and '-key' remain fallbacks for handshakes without HostName
9195 negotiation). If the unrecognized_name alert has to be sent, this by
9196 default is a warning; it becomes fatal with the '-servername_fatal'
9199 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9201 * Add AES and SSE2 assembly language support to VC++ build.
9205 * Mitigate attack on final subtraction in Montgomery reduction.
9209 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9210 (which previously caused an internal error).
9214 * Squeeze another 10% out of IGE mode when in != out.
9218 * AES IGE mode speedup.
9220 *Dean Gaudet (Google)*
9222 * Add the Korean symmetric 128-bit cipher SEED (see
9223 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9224 add SEED ciphersuites from RFC 4162:
9226 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9227 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9228 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9229 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9231 To minimize changes between patchlevels in the OpenSSL 0.9.8
9232 series, SEED remains excluded from compilation unless OpenSSL
9233 is configured with 'enable-seed'.
9235 *KISA, Bodo Moeller*
9237 * Mitigate branch prediction attacks, which can be practical if a
9238 single processor is shared, allowing a spy process to extract
9239 information. For detailed background information, see
9240 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9241 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9242 and Necessary Software Countermeasures"). The core of the change
9243 are new versions BN_div_no_branch() and
9244 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9245 respectively, which are slower, but avoid the security-relevant
9246 conditional branches. These are automatically called by BN_div()
9247 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9248 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9249 remove a conditional branch.
9251 BN_FLG_CONSTTIME is the new name for the previous
9252 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9253 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9254 in the exponent causes BN_mod_exp_mont() to use the alternative
9255 implementation in BN_mod_exp_mont_consttime().) The old name
9256 remains as a deprecated alias.
9258 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9259 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9260 constant-time implementations for more than just exponentiation.
9261 Here too the old name is kept as a deprecated alias.
9263 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9264 the BN_BLINDING structure gets an independent copy of the
9265 modulus. This means that the previous `BIGNUM *m` argument to
9266 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9267 essentially becomes `const BIGNUM *m`, although we can't actually
9268 change this in the header file before 0.9.9. It allows
9269 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9270 enable BN_FLG_CONSTTIME.
9272 *Matthew D Wood (Intel Corp)*
9274 * In the SSL/TLS server implementation, be strict about session ID
9275 context matching (which matters if an application uses a single
9276 external cache for different purposes). Previously,
9277 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9278 set. This did ensure strict client verification, but meant that,
9279 with applications using a single external cache for quite
9280 different requirements, clients could circumvent ciphersuite
9281 restrictions for a given session ID context by starting a session
9282 in a different context.
9286 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9287 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9288 authentication-only ciphersuites.
9292 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9293 not complete and could lead to a possible single byte overflow
9294 [CVE-2007-5135][] [Ben Laurie]
9296 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9298 * Since AES128 and AES256 (and similarly Camellia128 and
9299 Camellia256) share a single mask bit in the logic of
9300 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9301 kludge to work properly if AES128 is available and AES256 isn't
9302 (or if Camellia128 is available and Camellia256 isn't).
9306 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9307 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9308 When a point or a seed is encoded in a BIT STRING, we need to
9309 prevent the removal of trailing zero bits to get the proper DER
9310 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9311 of a NamedBitList, for which trailing 0 bits need to be removed.)
9315 * Have SSL/TLS server implementation tolerate "mismatched" record
9316 protocol version while receiving ClientHello even if the
9317 ClientHello is fragmented. (The server can't insist on the
9318 particular protocol version it has chosen before the ServerHello
9319 message has informed the client about his choice.)
9323 * Add RFC 3779 support.
9325 *Rob Austein for ARIN, Ben Laurie*
9327 * Load error codes if they are not already present instead of using a
9328 static variable. This allows them to be cleanly unloaded and reloaded.
9329 Improve header file function name parsing.
9333 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9334 or CAPABILITY handshake as required by RFCs.
9338 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9340 * Introduce limits to prevent malicious keys being able to
9341 cause a denial of service. [CVE-2006-2940][]
9343 *Steve Henson, Bodo Moeller*
9345 * Fix ASN.1 parsing of certain invalid structures that can result
9346 in a denial of service. [CVE-2006-2937][] [Steve Henson]
9348 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9349 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
9351 * Fix SSL client code which could crash if connecting to a
9352 malicious SSLv2 server. [CVE-2006-4343][]
9354 *Tavis Ormandy and Will Drewry, Google Security Team*
9356 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9357 match only those. Before that, "AES256-SHA" would be interpreted
9358 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9359 the same strength classification in 0.9.7h) as we currently only
9360 have a single AES bit in the ciphersuite description bitmap.
9361 That change, however, also applied to ciphersuite strings such as
9362 "RC4-MD5" that intentionally matched multiple ciphersuites --
9363 namely, SSL 2.0 ciphersuites in addition to the more common ones
9364 from SSL 3.0/TLS 1.0.
9366 So we change the selection algorithm again: Naming an explicit
9367 ciphersuite selects this one ciphersuite, and any other similar
9368 ciphersuite (same bitmap) from *other* protocol versions.
9369 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9370 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9372 Since SSL 2.0 does not have any ciphersuites for which the
9373 128/256 bit distinction would be relevant, this works for now.
9374 The proper fix will be to use different bits for AES128 and
9375 AES256, which would have avoided the problems from the beginning;
9376 however, bits are scarce, so we can only do this in a new release
9377 (not just a patchlevel) when we can change the SSL_CIPHER
9378 definition to split the single 'unsigned long mask' bitmap into
9379 multiple values to extend the available space.
9383 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9385 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9386 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
9388 * Add AES IGE and biIGE modes.
9392 * Change the Unix randomness entropy gathering to use poll() when
9393 possible instead of select(), since the latter has some
9394 undesirable limitations.
9396 *Darryl Miles via Richard Levitte and Bodo Moeller*
9398 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9399 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9400 cannot be implicitly activated as part of, e.g., the "AES" alias.
9401 However, please upgrade to OpenSSL 0.9.9[-dev] for
9402 non-experimental use of the ECC ciphersuites to get TLS extension
9403 support, which is required for curve and point format negotiation
9404 to avoid potential handshake problems.
9408 * Disable rogue ciphersuites:
9410 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9411 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9412 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9414 The latter two were purportedly from
9415 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9418 Also deactivate the remaining ciphersuites from
9419 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9420 unofficial, and the ID has long expired.
9424 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9425 dual-core machines) and other potential thread-safety issues.
9429 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9430 versions), which is now available for royalty-free use
9431 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9432 Also, add Camellia TLS ciphersuites from RFC 4132.
9434 To minimize changes between patchlevels in the OpenSSL 0.9.8
9435 series, Camellia remains excluded from compilation unless OpenSSL
9436 is configured with 'enable-camellia'.
9440 * Disable the padding bug check when compression is in use. The padding
9441 bug check assumes the first packet is of even length, this is not
9442 necessarily true if compression is enabled and can result in false
9443 positives causing handshake failure. The actual bug test is ancient
9444 code so it is hoped that implementations will either have fixed it by
9445 now or any which still have the bug do not support compression.
9449 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9451 * When applying a cipher rule check to see if string match is an explicit
9452 cipher suite and only match that one cipher suite if it is.
9456 * Link in manifests for VC++ if needed.
9458 *Austin Ziegler <halostatue@gmail.com>*
9460 * Update support for ECC-based TLS ciphersuites according to
9461 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9462 TLS extensions, which are supported starting with the 0.9.9
9463 branch, not in the OpenSSL 0.9.8 branch).
9467 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9468 opaque EVP_CIPHER_CTX handling.
9472 * Fixes and enhancements to zlib compression code. We now only use
9473 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9474 to conform with the standards mentioned here:
9475 <http://www.zlib.net/DLL_FAQ.txt>
9476 Static zlib linking now works on Windows and the new --with-zlib-include
9477 --with-zlib-lib options to Configure can be used to supply the location
9478 of the headers and library. Gracefully handle case where zlib library
9483 * Several fixes and enhancements to the OID generation code. The old code
9484 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9485 handle numbers larger than ULONG_MAX, truncated printing and had a
9486 non standard OBJ_obj2txt() behaviour.
9490 * Add support for building of engines under engine/ as shared libraries
9491 under VC++ build system.
9495 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9496 Hopefully, we will not see any false combination of paths any more.
9500 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9502 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9503 (part of SSL_OP_ALL). This option used to disable the
9504 countermeasure against man-in-the-middle protocol-version
9505 rollback in the SSL 2.0 server implementation, which is a bad
9506 idea. [CVE-2005-2969][]
9508 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9509 for Information Security, National Institute of Advanced Industrial
9510 Science and Technology [AIST], Japan)*
9512 * Add two function to clear and return the verify parameter flags.
9516 * Keep cipherlists sorted in the source instead of sorting them at
9517 runtime, thus removing the need for a lock.
9521 * Avoid some small subgroup attacks in Diffie-Hellman.
9523 *Nick Mathewson and Ben Laurie*
9525 * Add functions for well-known primes.
9529 * Extended Windows CE support.
9531 *Satoshi Nakamura and Andy Polyakov*
9533 * Initialize SSL_METHOD structures at compile time instead of during
9534 runtime, thus removing the need for a lock.
9538 * Make PKCS7_decrypt() work even if no certificate is supplied by
9539 attempting to decrypt each encrypted key in turn. Add support to
9544 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9546 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9549 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9553 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9554 key into the same file any more.
9558 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9562 * Add -utf8 command line and config file option to 'ca'.
9564 *Stefan <stf@udoma.org*
9566 * Removed the macro des_crypt(), as it seems to conflict with some
9567 libraries. Use DES_crypt().
9571 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9572 involves renaming the source and generated shared-libs for
9573 both. The engines will accept the corrected or legacy ids
9574 ('ncipher' and '4758_cca' respectively) when binding. NB,
9575 this only applies when building 'shared'.
9577 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9579 * Add attribute functions to EVP_PKEY structure. Modify
9580 PKCS12_create() to recognize a CSP name attribute and
9581 use it. Make -CSP option work again in pkcs12 utility.
9585 * Add new functionality to the bn blinding code:
9586 - automatic re-creation of the BN_BLINDING parameters after
9587 a fixed number of uses (currently 32)
9588 - add new function for parameter creation
9589 - introduce flags to control the update behaviour of the
9590 BN_BLINDING parameters
9591 - hide BN_BLINDING structure
9592 Add a second BN_BLINDING slot to the RSA structure to improve
9593 performance when a single RSA object is shared among several
9598 * Add support for DTLS.
9600 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9602 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9603 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9607 * Remove buggy and incomplete DH cert support from
9608 ssl/ssl_rsa.c and ssl/s3_both.c
9612 * Use SHA-1 instead of MD5 as the default digest algorithm for
9613 the apps/openssl applications.
9617 * Compile clean with "-Wall -Wmissing-prototypes
9618 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9619 DEBUG_SAFESTACK must also be set.
9623 * Change ./Configure so that certain algorithms can be disabled by default.
9624 The new counterpiece to "no-xxx" is "enable-xxx".
9626 The patented RC5 and MDC2 algorithms will now be disabled unless
9627 "enable-rc5" and "enable-mdc2", respectively, are specified.
9629 (IDEA remains enabled despite being patented. This is because IDEA
9630 is frequently required for interoperability, and there is no license
9631 fee for non-commercial use. As before, "no-idea" can be used to
9632 avoid this algorithm.)
9636 * Add processing of proxy certificates (see RFC 3820). This work was
9637 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9638 EGEE (Enabling Grids for E-science in Europe).
9642 * RC4 performance overhaul on modern architectures/implementations, such
9643 as Intel P4, IA-64 and AMD64.
9647 * New utility extract-section.pl. This can be used specify an alternative
9648 section number in a pod file instead of having to treat each file as
9649 a separate case in Makefile. This can be done by adding two lines to the
9652 =for comment openssl_section:XXX
9654 The blank line is mandatory.
9658 * New arguments -certform, -keyform and -pass for s_client and s_server
9659 to allow alternative format key and certificate files and passphrase
9664 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9665 update associated structures and add various utility functions.
9667 Add new policy related verify parameters, include policy checking in
9668 standard verify code. Enhance 'smime' application with extra parameters
9669 to support policy checking and print out.
9673 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9674 Nehemiah processors. These extensions support AES encryption in hardware
9675 as well as RNG (though RNG support is currently disabled).
9677 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9679 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9683 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9685 *Andy Polyakov and a number of other people*
9687 * Improved PowerPC platform support. Most notably BIGNUM assembler
9688 implementation contributed by IBM.
9690 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9692 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9693 exponent rather than 'unsigned long'. There is a corresponding change to
9694 the new 'rsa_keygen' element of the RSA_METHOD structure.
9696 *Jelte Jansen, Geoff Thorpe*
9698 * Functionality for creating the initial serial number file is now
9699 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9701 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9702 number file to 1, which is bound to cause problems. To avoid
9703 the problems while respecting compatibility between different 0.9.7
9704 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9705 CA.pl for serial number initialization. With the new release 0.9.8,
9706 we can fix the problem directly in the 'ca' utility.)
9710 * Reduced header interdependencies by declaring more opaque objects in
9711 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9712 give fewer recursive includes, which could break lazy source code - so
9713 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9714 developers should define this symbol when building and using openssl to
9715 ensure they track the recommended behaviour, interfaces, [etc], but
9716 backwards-compatible behaviour prevails when this isn't defined.
9720 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9724 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9725 This will generate a random key of the appropriate length based on the
9726 cipher context. The EVP_CIPHER can provide its own random key generation
9727 routine to support keys of a specific form. This is used in the des and
9728 3des routines to generate a key of the correct parity. Update S/MIME
9729 code to use new functions and hence generate correct parity DES keys.
9730 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9731 valid (weak or incorrect parity).
9735 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9736 as looking them up. This is useful when the verified structure may contain
9737 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9738 present unless the new PKCS7_NO_CRL flag is asserted.
9742 * Extend ASN1 oid configuration module. It now additionally accepts the
9745 shortName = some long name, 1.2.3.4
9749 * Reimplemented the BN_CTX implementation. There is now no more static
9750 limitation on the number of variables it can handle nor the depth of the
9751 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9752 information can now expand as required, and rather than having a single
9753 static array of bignums, BN_CTX now uses a linked-list of such arrays
9754 allowing it to expand on demand whilst maintaining the usefulness of
9755 BN_CTX's "bundling".
9759 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9760 to allow all RSA operations to function using a single BN_CTX.
9764 * Preliminary support for certificate policy evaluation and checking. This
9765 is initially intended to pass the tests outlined in "Conformance Testing
9766 of Relying Party Client Certificate Path Processing Logic" v1.07.
9770 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9771 remained unused and not that useful. A variety of other little bignum
9772 tweaks and fixes have also been made continuing on from the audit (see
9777 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9778 associated ASN1, EVP and SSL functions and old ASN1 macros.
9782 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9783 and this should never fail. So the return value from the use of
9784 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9785 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9789 * BN_CTX_get() should return zero-valued bignums, providing the same
9790 initialised value as BN_new().
9792 *Geoff Thorpe, suggested by Ulf Möller*
9794 * Support for inhibitAnyPolicy certificate extension.
9798 * An audit of the BIGNUM code is underway, for which debugging code is
9799 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9800 is considered valid when processing BIGNUMs, and causes execution to
9801 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9802 further steps are taken to deliberately pollute unused data in BIGNUM
9803 structures to try and expose faulty code further on. For now, openssl will
9804 (in its default mode of operation) continue to tolerate the inconsistent
9805 forms that it has tolerated in the past, but authors and packagers should
9806 consider trying openssl and their own applications when compiled with
9807 these debugging symbols defined. It will help highlight potential bugs in
9808 their own code, and will improve the test coverage for OpenSSL itself. At
9809 some point, these tighter rules will become openssl's default to improve
9810 maintainability, though the assert()s and other overheads will remain only
9811 in debugging configurations. See bn.h for more details.
9813 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9815 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9816 that can only be obtained through BN_CTX_new() (which implicitly
9817 initialises it). The presence of this function only made it possible
9818 to overwrite an existing structure (and cause memory leaks).
9822 * Because of the callback-based approach for implementing LHASH as a
9823 template type, lh_insert() adds opaque objects to hash-tables and
9824 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9825 to clean up those corresponding objects before destroying the hash table
9826 (and losing the object pointers). So some over-zealous constifications in
9827 LHASH have been relaxed so that lh_insert() does not take (nor store) the
9828 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
9829 prototyped to have "const" restrictions on the object pointers they are
9830 given (and so aren't required to cast them away any more).
9834 * The tmdiff.h API was so ugly and minimal that our own timing utility
9835 (speed) prefers to use its own implementation. The two implementations
9836 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
9837 its object type properly exposed (MS_TM) instead of casting to/from
9838 `char *`. This may still change yet if someone realises MS_TM and
9840 aren't necessarily the greatest nomenclatures - but this is what was used
9841 internally to the implementation so I've used that for now.
9845 * Ensure that deprecated functions do not get compiled when
9846 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
9847 the self-tests were still using deprecated key-generation functions so
9848 these have been updated also.
9852 * Reorganise PKCS#7 code to separate the digest location functionality
9853 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
9854 New function PKCS7_set_digest() to set the digest type for PKCS#7
9855 digestedData type. Add additional code to correctly generate the
9856 digestedData type and add support for this type in PKCS7 initialization
9861 * New function PKCS7_set0_type_other() this initializes a PKCS7
9862 structure of type "other".
9866 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
9867 sure the loop does correctly stop and breaking ("division by zero")
9868 modulus operations are not performed. The (pre-generated) prime
9869 table crypto/bn/bn_prime.h was already correct, but it could not be
9870 re-generated on some platforms because of the "division by zero"
9871 situation in the script.
9873 *Ralf S. Engelschall*
9875 * Update support for ECC-based TLS ciphersuites according to
9876 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
9877 SHA-1 now is only used for "small" curves (where the
9878 representation of a field element takes up to 24 bytes); for
9879 larger curves, the field element resulting from ECDH is directly
9880 used as premaster secret.
9882 *Douglas Stebila (Sun Microsystems Laboratories)*
9884 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
9885 curve secp160r1 to the tests.
9887 *Douglas Stebila (Sun Microsystems Laboratories)*
9889 * Add the possibility to load symbols globally with DSO.
9891 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
9893 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
9894 control of the error stack.
9898 * Add support for STORE in ENGINE.
9902 * Add the STORE type. The intention is to provide a common interface
9903 to certificate and key stores, be they simple file-based stores, or
9904 HSM-type store, or LDAP stores, or...
9905 NOTE: The code is currently UNTESTED and isn't really used anywhere.
9909 * Add a generic structure called OPENSSL_ITEM. This can be used to
9910 pass a list of arguments to any function as well as provide a way
9911 for a function to pass data back to the caller.
9915 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
9916 works like BUF_strdup() but can be used to duplicate a portion of
9917 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
9922 * Add the function sk_find_ex() which works like sk_find(), but will
9923 return an index to an element even if an exact match couldn't be
9924 found. The index is guaranteed to point at the element where the
9925 searched-for key would be inserted to preserve sorting order.
9929 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
9930 takes an extra flags argument for optional functionality. Currently,
9931 the following flags are defined:
9933 OBJ_BSEARCH_VALUE_ON_NOMATCH
9934 This one gets OBJ_bsearch_ex() to return a pointer to the first
9935 element where the comparing function returns a negative or zero
9938 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
9939 This one gets OBJ_bsearch_ex() to return a pointer to the first
9940 element where the comparing function returns zero. This is useful
9941 if there are more than one element where the comparing function
9946 * Make it possible to create self-signed certificates with 'openssl ca'
9947 in such a way that the self-signed certificate becomes part of the
9948 CA database and uses the same mechanisms for serial number generation
9949 as all other certificate signing. The new flag '-selfsign' enables
9950 this functionality. Adapt CA.sh and CA.pl.in.
9954 * Add functionality to check the public key of a certificate request
9955 against a given private. This is useful to check that a certificate
9956 request can be signed by that key (self-signing).
9960 * Make it possible to have multiple active certificates with the same
9961 subject in the CA index file. This is done only if the keyword
9962 'unique_subject' is set to 'no' in the main CA section (default
9963 if 'CA_default') of the configuration file. The value is saved
9964 with the database itself in a separate index attribute file,
9965 named like the index file with '.attr' appended to the name.
9969 * Generate multi-valued AVAs using '+' notation in config files for
9974 * Support for nameConstraints certificate extension.
9978 * Support for policyConstraints certificate extension.
9982 * Support for policyMappings certificate extension.
9986 * Make sure the default DSA_METHOD implementation only uses its
9987 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
9988 and change its own handlers to be NULL so as to remove unnecessary
9989 indirection. This lets alternative implementations fallback to the
9990 default implementation more easily.
9994 * Support for directoryName in GeneralName related extensions
9999 * Make it possible to link applications using Makefile.shared.
10000 Make that possible even when linking against static libraries!
10004 * Support for single pass processing for S/MIME signing. This now
10005 means that S/MIME signing can be done from a pipe, in addition
10006 cleartext signing (multipart/signed type) is effectively streaming
10007 and the signed data does not need to be all held in memory.
10009 This is done with a new flag PKCS7_STREAM. When this flag is set
10010 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10011 is done after the data is output (and digests calculated) in
10012 SMIME_write_PKCS7().
10016 * Add full support for -rpath/-R, both in shared libraries and
10017 applications, at least on the platforms where it's known how
10022 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10023 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10024 will now compute a table of multiples of the generator that
10025 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10026 faster (notably in the case of a single point multiplication,
10027 scalar * generator).
10029 *Nils Larsch, Bodo Moeller*
10031 * IPv6 support for certificate extensions. The various extensions
10032 which use the IP:a.b.c.d can now take IPv6 addresses using the
10033 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10038 * Added an ENGINE that implements RSA by performing private key
10039 exponentiations with the GMP library. The conversions to and from
10040 GMP's mpz_t format aren't optimised nor are any montgomery forms
10041 cached, and on x86 it appears OpenSSL's own performance has caught up.
10042 However there are likely to be other architectures where GMP could
10043 provide a boost. This ENGINE is not built in by default, but it can be
10044 specified at Configure time and should be accompanied by the necessary
10045 linker additions, eg;
10046 ./config -DOPENSSL_USE_GMP -lgmp
10050 * "openssl engine" will not display ENGINE/DSO load failure errors when
10051 testing availability of engines with "-t" - the old behaviour is
10052 produced by increasing the feature's verbosity with "-tt".
10056 * ECDSA routines: under certain error conditions uninitialized BN objects
10057 could be freed. Solution: make sure initialization is performed early
10058 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10063 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10064 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10065 software implementations. For DSA and DH, parameter generation can
10066 also be overridden by providing the appropriate method callbacks.
10070 * Change the "progress" mechanism used in key-generation and
10071 primality testing to functions that take a new BN_GENCB pointer in
10072 place of callback/argument pairs. The new API functions have `_ex`
10073 postfixes and the older functions are reimplemented as wrappers for
10074 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10075 declarations of the old functions to help (graceful) attempts to
10076 migrate to the new functions. Also, the new key-generation API
10077 functions operate on a caller-supplied key-structure and return
10078 success/failure rather than returning a key or NULL - this is to
10079 help make "keygen" another member function of RSA_METHOD etc.
10081 Example for using the new callback interface:
10083 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10084 void *my_arg = ...;
10087 BN_GENCB_set(&my_cb, my_callback, my_arg);
10089 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10090 /* For the meaning of a, b in calls to my_callback(), see the
10091 * documentation of the function that calls the callback.
10092 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10093 * my_callback should return 1 if it wants BN_is_prime_ex()
10094 * to continue, or 0 to stop.
10099 * Change the ZLIB compression method to be stateful, and make it
10100 available to TLS with the number defined in
10101 draft-ietf-tls-compression-04.txt.
10105 * Add the ASN.1 structures and functions for CertificatePair, which
10106 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10108 CertificatePair ::= SEQUENCE {
10109 forward [0] Certificate OPTIONAL,
10110 reverse [1] Certificate OPTIONAL,
10111 -- at least one of the pair shall be present -- }
10113 Also implement the PEM functions to read and write certificate
10114 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10116 This needed to be defined, mostly for the sake of the LDAP
10117 attribute crossCertificatePair, but may prove useful elsewhere as
10122 * Make it possible to inhibit symlinking of shared libraries in
10123 Makefile.shared, for Cygwin's sake.
10127 * Extend the BIGNUM API by creating a function
10128 void BN_set_negative(BIGNUM *a, int neg);
10129 and a macro that behave like
10130 int BN_is_negative(const BIGNUM *a);
10132 to avoid the need to access 'a->neg' directly in applications.
10136 * Implement fast modular reduction for pseudo-Mersenne primes
10137 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10138 EC_GROUP_new_curve_GFp() will now automatically use this
10141 *Nils Larsch <nla@trustcenter.de>*
10143 * Add new lock type (CRYPTO_LOCK_BN).
10147 * Change the ENGINE framework to automatically load engines
10148 dynamically from specific directories unless they could be
10149 found to already be built in or loaded. Move all the
10150 current engines except for the cryptodev one to a new
10151 directory engines/.
10152 The engines in engines/ are built as shared libraries if
10153 the "shared" options was given to ./Configure or ./config.
10154 Otherwise, they are inserted in libcrypto.a.
10155 /usr/local/ssl/engines is the default directory for dynamic
10156 engines, but that can be overridden at configure time through
10157 the usual use of --prefix and/or --openssldir, and at run
10158 time with the environment variable OPENSSL_ENGINES.
10160 *Geoff Thorpe and Richard Levitte*
10162 * Add Makefile.shared, a helper makefile to build shared
10163 libraries. Adapt Makefile.org.
10167 * Add version info to Win32 DLLs.
10169 *Peter 'Luna' Runestig" <peter@runestig.com>*
10171 * Add new 'medium level' PKCS#12 API. Certificates and keys
10172 can be added using this API to created arbitrary PKCS#12
10173 files while avoiding the low level API.
10175 New options to PKCS12_create(), key or cert can be NULL and
10176 will then be omitted from the output file. The encryption
10177 algorithm NIDs can be set to -1 for no encryption, the mac
10178 iteration count can be set to 0 to omit the mac.
10180 Enhance pkcs12 utility by making the -nokeys and -nocerts
10181 options work when creating a PKCS#12 file. New option -nomac
10182 to omit the mac, NONE can be set for an encryption algorithm.
10183 New code is modified to use the enhanced PKCS12_create()
10184 instead of the low level API.
10188 * Extend ASN1 encoder to support indefinite length constructed
10189 encoding. This can output sequences tags and octet strings in
10190 this form. Modify pk7_asn1.c to support indefinite length
10191 encoding. This is experimental and needs additional code to
10192 be useful, such as an ASN1 bio and some enhanced streaming
10195 Extend template encode functionality so that tagging is passed
10196 down to the template encoder.
10200 * Let 'openssl req' fail if an argument to '-newkey' is not
10201 recognized instead of using RSA as a default.
10205 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10206 As these are not official, they are not included in "ALL";
10207 the "ECCdraft" ciphersuite group alias can be used to select them.
10209 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10211 * Add ECDH engine support.
10213 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10215 * Add ECDH in new directory crypto/ecdh/.
10217 *Douglas Stebila (Sun Microsystems Laboratories)*
10219 * Let BN_rand_range() abort with an error after 100 iterations
10220 without success (which indicates a broken PRNG).
10224 * Change BN_mod_sqrt() so that it verifies that the input value
10225 is really the square of the return value. (Previously,
10226 BN_mod_sqrt would show GIGO behaviour.)
10230 * Add named elliptic curves over binary fields from X9.62, SECG,
10231 and WAP/WTLS; add OIDs that were still missing.
10233 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10235 * Extend the EC library for elliptic curves over binary fields
10236 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10239 EC_GF2m_simple_method
10243 EC_GROUP_new_curve_GF2m
10244 EC_GROUP_set_curve_GF2m
10245 EC_GROUP_get_curve_GF2m
10246 EC_POINT_set_affine_coordinates_GF2m
10247 EC_POINT_get_affine_coordinates_GF2m
10248 EC_POINT_set_compressed_coordinates_GF2m
10250 Point compression for binary fields is disabled by default for
10251 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10254 As binary polynomials are represented as BIGNUMs, various members
10255 of the EC_GROUP and EC_POINT data structures can be shared
10256 between the implementations for prime fields and binary fields;
10257 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10258 are essentially identical to their `..._GFp` counterparts.
10259 (For simplicity, the `..._GFp` prefix has been dropped from
10260 various internal method names.)
10262 An internal 'field_div' method (similar to 'field_mul' and
10263 'field_sqr') has been added; this is used only for binary fields.
10265 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10267 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10268 through methods ('mul', 'precompute_mult').
10270 The generic implementations (now internally called 'ec_wNAF_mul'
10271 and 'ec_wNAF_precomputed_mult') remain the default if these
10272 methods are undefined.
10274 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10276 * New function EC_GROUP_get_degree, which is defined through
10277 EC_METHOD. For curves over prime fields, this returns the bit
10278 length of the modulus.
10280 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10282 * New functions EC_GROUP_dup, EC_POINT_dup.
10283 (These simply call ..._new and ..._copy).
10285 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10287 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10288 Polynomials are represented as BIGNUMs (where the sign bit is not
10289 used) in the following functions [macros]:
10292 BN_GF2m_sub [= BN_GF2m_add]
10293 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10294 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10295 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10297 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10298 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10299 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10300 BN_GF2m_cmp [= BN_ucmp]
10302 (Note that only the 'mod' functions are actually for fields GF(2^m).
10303 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10305 For some functions, an the irreducible polynomial defining a
10306 field can be given as an 'unsigned int[]' with strictly
10307 decreasing elements giving the indices of those bits that are set;
10308 i.e., p[] represents the polynomial
10309 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10311 p[0] > p[1] > ... > p[k] = 0.
10312 This applies to the following functions:
10315 BN_GF2m_mod_mul_arr
10316 BN_GF2m_mod_sqr_arr
10317 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10318 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10319 BN_GF2m_mod_exp_arr
10320 BN_GF2m_mod_sqrt_arr
10321 BN_GF2m_mod_solve_quad_arr
10325 Conversion can be performed by the following functions:
10330 bntest.c has additional tests for binary polynomial arithmetic.
10332 Two implementations for BN_GF2m_mod_div() are available.
10333 The default algorithm simply uses BN_GF2m_mod_inv() and
10334 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10335 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10336 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10338 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10340 * Add new error code 'ERR_R_DISABLED' that can be used when some
10341 functionality is disabled at compile-time.
10343 *Douglas Stebila <douglas.stebila@sun.com>*
10345 * Change default behaviour of 'openssl asn1parse' so that more
10346 information is visible when viewing, e.g., a certificate:
10348 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10349 mode the content of non-printable OCTET STRINGs is output in a
10350 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10351 avoid the appearance of a printable string.
10353 *Nils Larsch <nla@trustcenter.de>*
10355 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10357 EC_GROUP_set_asn1_flag()
10358 EC_GROUP_get_asn1_flag()
10359 EC_GROUP_set_point_conversion_form()
10360 EC_GROUP_get_point_conversion_form()
10361 These control ASN1 encoding details:
10362 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10363 has been set to OPENSSL_EC_NAMED_CURVE.
10364 - Points are encoded in uncompressed form by default; options for
10365 asn1_for are as for point2oct, namely
10366 POINT_CONVERSION_COMPRESSED
10367 POINT_CONVERSION_UNCOMPRESSED
10368 POINT_CONVERSION_HYBRID
10370 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10372 EC_GROUP_set_seed()
10373 EC_GROUP_get0_seed()
10374 EC_GROUP_get_seed_len()
10375 This is used only for ASN1 purposes (so far).
10377 *Nils Larsch <nla@trustcenter.de>*
10379 * Add 'field_type' member to EC_METHOD, which holds the NID
10380 of the appropriate field type OID. The new function
10381 EC_METHOD_get_field_type() returns this value.
10383 *Nils Larsch <nla@trustcenter.de>*
10386 EC_POINT_point2bn()
10387 EC_POINT_bn2point()
10388 EC_POINT_point2hex()
10389 EC_POINT_hex2point()
10390 providing useful interfaces to EC_POINT_point2oct() and
10391 EC_POINT_oct2point().
10393 *Nils Larsch <nla@trustcenter.de>*
10395 * Change internals of the EC library so that the functions
10396 EC_GROUP_set_generator()
10397 EC_GROUP_get_generator()
10398 EC_GROUP_get_order()
10399 EC_GROUP_get_cofactor()
10400 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10401 to methods, which would lead to unnecessary code duplication when
10402 adding different types of curves.
10404 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10406 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10407 arithmetic, and such that modified wNAFs are generated
10408 (which avoid length expansion in many cases).
10412 * Add a function EC_GROUP_check_discriminant() (defined via
10413 EC_METHOD) that verifies that the curve discriminant is non-zero.
10415 Add a function EC_GROUP_check() that makes some sanity tests
10416 on a EC_GROUP, its generator and order. This includes
10417 EC_GROUP_check_discriminant().
10419 *Nils Larsch <nla@trustcenter.de>*
10421 * Add ECDSA in new directory crypto/ecdsa/.
10423 Add applications 'openssl ecparam' and 'openssl ecdsa'
10424 (these are based on 'openssl dsaparam' and 'openssl dsa').
10426 ECDSA support is also included in various other files across the
10427 library. Most notably,
10428 - 'openssl req' now has a '-newkey ecdsa:file' option;
10429 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10430 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10431 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10432 them suitable for ECDSA where domain parameters must be
10433 extracted before the specific public key;
10434 - ECDSA engine support has been added.
10436 *Nils Larsch <nla@trustcenter.de>*
10438 * Include some named elliptic curves, and add OIDs from X9.62,
10439 SECG, and WAP/WTLS. Each curve can be obtained from the new
10441 EC_GROUP_new_by_curve_name(),
10442 and the list of available named curves can be obtained with
10443 EC_get_builtin_curves().
10444 Also add a 'curve_name' member to EC_GROUP objects, which can be
10446 EC_GROUP_set_curve_name()
10447 EC_GROUP_get_curve_name()
10449 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10451 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10452 was actually never needed) and in BN_mul(). The removal in BN_mul()
10453 required a small change in bn_mul_part_recursive() and the addition
10454 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10455 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10456 bn_sub_words() and bn_add_words() except they take arrays with
10461 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10463 * Cleanse PEM buffers before freeing them since they may contain
10466 *Benjamin Bennett <ben@psc.edu>*
10468 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10469 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10470 authentication-only ciphersuites.
10474 * Since AES128 and AES256 share a single mask bit in the logic of
10475 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10476 kludge to work properly if AES128 is available and AES256 isn't.
10480 * Expand security boundary to match 1.1.1 module.
10484 * Remove redundant features: hash file source, editing of test vectors
10485 modify fipsld to use external fips_premain.c signature.
10489 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10490 run algorithm test programs.
10494 * Make algorithm test programs more tolerant of whitespace.
10498 * Have SSL/TLS server implementation tolerate "mismatched" record
10499 protocol version while receiving ClientHello even if the
10500 ClientHello is fragmented. (The server can't insist on the
10501 particular protocol version it has chosen before the ServerHello
10502 message has informed the client about his choice.)
10506 * Load error codes if they are not already present instead of using a
10507 static variable. This allows them to be cleanly unloaded and reloaded.
10511 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10513 * Introduce limits to prevent malicious keys being able to
10514 cause a denial of service. [CVE-2006-2940][]
10516 *Steve Henson, Bodo Moeller*
10518 * Fix ASN.1 parsing of certain invalid structures that can result
10519 in a denial of service. [CVE-2006-2937][] [Steve Henson]
10521 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10522 [CVE-2006-3738][] [Tavis Ormandy and Will Drewry, Google Security Team]
10524 * Fix SSL client code which could crash if connecting to a
10525 malicious SSLv2 server. [CVE-2006-4343][]
10527 *Tavis Ormandy and Will Drewry, Google Security Team*
10529 * Change ciphersuite string processing so that an explicit
10530 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10531 will no longer include "AES128-SHA"), and any other similar
10532 ciphersuite (same bitmap) from *other* protocol versions (so that
10533 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10534 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10535 changes from 0.9.8b and 0.9.8d.
10539 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10541 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10542 [CVE-2006-4339][] [Ben Laurie and Google Security Team]
10544 * Change the Unix randomness entropy gathering to use poll() when
10545 possible instead of select(), since the latter has some
10546 undesirable limitations.
10548 *Darryl Miles via Richard Levitte and Bodo Moeller*
10550 * Disable rogue ciphersuites:
10552 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10553 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10554 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10556 The latter two were purportedly from
10557 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10560 Also deactivate the remaining ciphersuites from
10561 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10562 unofficial, and the ID has long expired.
10566 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10567 dual-core machines) and other potential thread-safety issues.
10571 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10573 * Adapt fipsld and the build system to link against the validated FIPS
10574 module in FIPS mode.
10578 * Fixes for VC++ 2005 build under Windows.
10582 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10583 from a Windows bash shell such as MSYS. It is autodetected from the
10584 "config" script when run from a VC++ environment. Modify standard VC++
10585 build to use fipscanister.o from the GNU make build.
10589 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10591 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10592 The value now differs depending on if you build for FIPS or not.
10593 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10594 safely run with a non-FIPSed libcrypto, as it may crash because of
10595 the difference induced by this change.
10599 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10601 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10602 (part of SSL_OP_ALL). This option used to disable the
10603 countermeasure against man-in-the-middle protocol-version
10604 rollback in the SSL 2.0 server implementation, which is a bad
10605 idea. [CVE-2005-2969][]
10607 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10608 for Information Security, National Institute of Advanced Industrial
10609 Science and Technology [AIST, Japan)]*
10611 * Minimal support for X9.31 signatures and PSS padding modes. This is
10612 mainly for FIPS compliance and not fully integrated at this stage.
10616 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10617 the exponentiation using a fixed-length exponent. (Otherwise,
10618 the information leaked through timing could expose the secret key
10619 after many signatures; cf. Bleichenbacher's attack on DSA with
10624 * Make a new fixed-window mod_exp implementation the default for
10625 RSA, DSA, and DH private-key operations so that the sequence of
10626 squares and multiplies and the memory access pattern are
10627 independent of the particular secret key. This will mitigate
10628 cache-timing and potential related attacks.
10630 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10631 and this is automatically used by BN_mod_exp_mont() if the new flag
10632 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10633 will use this BN flag for private exponents unless the flag
10634 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10635 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10637 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10639 * Change the client implementation for SSLv23_method() and
10640 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10641 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10642 (Previously, the SSL 2.0 backwards compatible Client Hello
10643 message format would be used even with SSL_OP_NO_SSLv2.)
10647 * Add support for smime-type MIME parameter in S/MIME messages which some
10652 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10653 a threadsafe manner. Modify rsa code to use new function and add calls
10654 to dsa and dh code (which had race conditions before).
10658 * Include the fixed error library code in the C error file definitions
10659 instead of fixing them up at runtime. This keeps the error code
10660 structures constant.
10664 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10666 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10669 * Fixes for newer kerberos headers. NB: the casts are needed because
10670 the 'length' field is signed on one version and unsigned on another
10671 with no (?) obvious way to tell the difference, without these VC++
10672 complains. Also the "definition" of FAR (blank) is no longer included
10673 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10674 some needed definitions.
10678 * Undo Cygwin change.
10682 * Added support for proxy certificates according to RFC 3820.
10683 Because they may be a security thread to unaware applications,
10684 they must be explicitly allowed in run-time. See
10685 docs/HOWTO/proxy_certificates.txt for further information.
10689 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10691 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10692 server and client random values. Previously
10693 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10694 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10696 This change has negligible security impact because:
10698 1. Server and client random values still have 24 bytes of pseudo random
10701 2. Server and client random values are sent in the clear in the initial
10704 3. The master secret is derived using the premaster secret (48 bytes in
10705 size for static RSA ciphersuites) as well as client server and random
10708 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10711 *Stephen Henson, reported by UK NISCC*
10713 * Use Windows randomness collection on Cygwin.
10717 * Fix hang in EGD/PRNGD query when communication socket is closed
10718 prematurely by EGD/PRNGD.
10720 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10722 * Prompt for pass phrases when appropriate for PKCS12 input format.
10726 * Back-port of selected performance improvements from development
10727 branch, as well as improved support for PowerPC platforms.
10731 * Add lots of checks for memory allocation failure, error codes to indicate
10732 failure and freeing up memory if a failure occurs.
10734 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10736 * Add new -passin argument to dgst.
10740 * Perform some character comparisons of different types in X509_NAME_cmp:
10741 this is needed for some certificates that re-encode DNs into UTF8Strings
10742 (in violation of RFC3280) and can't or won't issue name rollover
10747 * Make an explicit check during certificate validation to see that
10748 the CA setting in each certificate on the chain is correct. As a
10749 side effect always do the following basic checks on extensions,
10750 not just when there's an associated purpose to the check:
10752 - if there is an unhandled critical extension (unless the user
10753 has chosen to ignore this fault)
10754 - if the path length has been exceeded (if one is set at all)
10755 - that certain extensions fit the associated purpose (if one has
10760 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10762 * Avoid a race condition when CRLs are checked in a multi threaded
10763 environment. This would happen due to the reordering of the revoked
10764 entries during signature checking and serial number lookup. Now the
10765 encoding is cached and the serial number sort performed under a lock.
10766 Add new STACK function sk_is_sorted().
10770 * Add Delta CRL to the extension code.
10774 * Various fixes to s3_pkt.c so alerts are sent properly.
10776 *David Holmes <d.holmes@f5.com>*
10778 * Reduce the chances of duplicate issuer name and serial numbers (in
10779 violation of RFC3280) using the OpenSSL certificate creation utilities.
10780 This is done by creating a random 64 bit value for the initial serial
10781 number when a serial number file is created or when a self signed
10782 certificate is created using 'openssl req -x509'. The initial serial
10783 number file is created using 'openssl x509 -next_serial' in CA.pl
10784 rather than being initialized to 1.
10788 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10790 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10791 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
10793 *Joe Orton, Steve Henson*
10795 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10798 *Joe Orton, Steve Henson*
10800 * Make it possible to have multiple active certificates with the same
10801 subject in the CA index file. This is done only if the keyword
10802 'unique_subject' is set to 'no' in the main CA section (default
10803 if 'CA_default') of the configuration file. The value is saved
10804 with the database itself in a separate index attribute file,
10805 named like the index file with '.attr' appended to the name.
10809 * X509 verify fixes. Disable broken certificate workarounds when
10810 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10811 keyUsage extension present. Don't accept CRLs with unhandled critical
10812 extensions: since verify currently doesn't process CRL extensions this
10813 rejects a CRL with *any* critical extensions. Add new verify error codes
10818 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10819 A clarification of RFC2560 will require the use of OCTET STRINGs and
10820 some implementations cannot handle the current raw format. Since OpenSSL
10821 copies and compares OCSP nonces as opaque blobs without any attempt at
10822 parsing them this should not create any compatibility issues.
10826 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
10827 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
10828 this HMAC (and other) operations are several times slower than OpenSSL
10833 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
10835 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
10837 * Use the correct content when signing type "other".
10841 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
10843 * Fix various bugs revealed by running the NISCC test suite:
10845 Stop out of bounds reads in the ASN1 code when presented with
10846 invalid tags (CVE-2003-0543 and CVE-2003-0544).
10848 Free up ASN1_TYPE correctly if ANY type is invalid [CVE-2003-0545][].
10850 If verify callback ignores invalid public key errors don't try to check
10851 certificate signature with the NULL public key.
10855 * New -ignore_err option in ocsp application to stop the server
10856 exiting on the first error in a request.
10860 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
10861 if the server requested one: as stated in TLS 1.0 and SSL 3.0
10866 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
10867 extra data after the compression methods not only for TLS 1.0
10868 but also for SSL 3.0 (as required by the specification).
10870 *Bodo Moeller; problem pointed out by Matthias Loepfe*
10872 * Change X509_certificate_type() to mark the key as exported/exportable
10873 when it's 512 *bits* long, not 512 bytes.
10877 * Change AES_cbc_encrypt() so it outputs exact multiple of
10878 blocks during encryption.
10882 * Various fixes to base64 BIO and non blocking I/O. On write
10883 flushes were not handled properly if the BIO retried. On read
10884 data was not being buffered properly and had various logic bugs.
10885 This also affects blocking I/O when the data being decoded is a
10890 * Various S/MIME bugfixes and compatibility changes:
10891 output correct application/pkcs7 MIME type if
10892 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
10893 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
10894 of files as .eml work). Correctly handle very long lines in MIME
10899 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
10901 * Countermeasure against the Klima-Pokorny-Rosa extension of
10902 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
10903 a protocol version number mismatch like a decryption error
10904 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
10908 * Turn on RSA blinding by default in the default implementation
10909 to avoid a timing attack. Applications that don't want it can call
10910 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
10911 They would be ill-advised to do so in most cases.
10913 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
10915 * Change RSA blinding code so that it works when the PRNG is not
10916 seeded (in this case, the secret RSA exponent is abused as
10917 an unpredictable seed -- if it is not unpredictable, there
10918 is no point in blinding anyway). Make RSA blinding thread-safe
10919 by remembering the creator's thread ID in rsa->blinding and
10920 having all other threads use local one-time blinding factors
10921 (this requires more computation than sharing rsa->blinding, but
10922 avoids excessive locking; and if an RSA object is not shared
10923 between threads, blinding will still be very fast).
10927 * Fixed a typo bug that would cause ENGINE_set_default() to set an
10928 ENGINE as defaults for all supported algorithms irrespective of
10929 the 'flags' parameter. 'flags' is now honoured, so applications
10930 should make sure they are passing it correctly.
10934 * Target "mingw" now allows native Windows code to be generated in
10935 the Cygwin environment as well as with the MinGW compiler.
10939 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
10941 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
10942 via timing by performing a MAC computation even if incorrect
10943 block cipher padding has been found. This is a countermeasure
10944 against active attacks where the attacker has to distinguish
10945 between bad padding and a MAC verification error. [CVE-2003-0078][]
10947 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
10948 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
10949 Martin Vuagnoux (EPFL, Ilion)*
10951 * Make the no-err option work as intended. The intention with no-err
10952 is not to have the whole error stack handling routines removed from
10953 libcrypto, it's only intended to remove all the function name and
10954 reason texts, thereby removing some of the footprint that may not
10955 be interesting if those errors aren't displayed anyway.
10957 NOTE: it's still possible for any application or module to have its
10958 own set of error texts inserted. The routines are there, just not
10959 used by default when no-err is given.
10963 * Add support for FreeBSD on IA64.
10965 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
10967 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
10968 Kerberos function mit_des_cbc_cksum(). Before this change,
10969 the value returned by DES_cbc_cksum() was like the one from
10970 mit_des_cbc_cksum(), except the bytes were swapped.
10972 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
10974 * Allow an application to disable the automatic SSL chain building.
10975 Before this a rather primitive chain build was always performed in
10976 ssl3_output_cert_chain(): an application had no way to send the
10977 correct chain if the automatic operation produced an incorrect result.
10979 Now the chain builder is disabled if either:
10981 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
10983 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
10985 The reasoning behind this is that an application would not want the
10986 auto chain building to take place if extra chain certificates are
10987 present and it might also want a means of sending no additional
10988 certificates (for example the chain has two certificates and the
10993 * Add the possibility to build without the ENGINE framework.
10995 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
10997 * Under Win32 gmtime() can return NULL: check return value in
10998 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11002 * DSA routines: under certain error conditions uninitialized BN objects
11003 could be freed. Solution: make sure initialization is performed early
11004 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11005 Nils Larsch <nla@trustcenter.de> via PR#459)
11009 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11010 checked on reconnect on the client side, therefore session resumption
11011 could still fail with a "ssl session id is different" error. This
11012 behaviour is masked when SSL_OP_ALL is used due to
11013 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11014 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11015 followup to PR #377.
11019 * IA-32 assembler support enhancements: unified ELF targets, support
11020 for SCO/Caldera platforms, fix for Cygwin shared build.
11024 * Add support for FreeBSD on sparc64. As a consequence, support for
11025 FreeBSD on non-x86 processors is separate from x86 processors on
11026 the config script, much like the NetBSD support.
11028 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11030 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11032 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11035 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11036 code (06) was taken as the first octet of the session ID and the last
11037 octet was ignored consequently. As a result SSLv2 client side session
11038 caching could not have worked due to the session ID mismatch between
11040 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11045 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11046 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11051 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11052 seems that in spite of existing for more than a year, many application
11053 author have done nothing to provide the necessary callbacks, which
11054 means that this particular engine will not work properly anywhere.
11055 This is a very unfortunate situation which forces us, in the name
11056 of usability, to give the hw_ncipher.c a static lock, which is part
11058 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11059 appear in 0.9.8 or later. We EXPECT application authors to have
11060 dealt properly with this when 0.9.8 is released (unless we actually
11061 make such changes in the libcrypto locking code that changes will
11062 have to be made anyway).
11066 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11067 octets have been read, EOF or an error occurs. Without this change
11068 some truncated ASN1 structures will not produce an error.
11072 * Disable Heimdal support, since it hasn't been fully implemented.
11073 Still give the possibility to force the use of Heimdal, but with
11074 warnings and a request that patches get sent to openssl-dev.
11078 * Add the VC-CE target, introduce the WINCE sysname, and add
11079 INSTALL.WCE and appropriate conditionals to make it build.
11081 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11083 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11084 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11085 edit numbers of the version.
11087 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11089 * Introduce safe string copy and catenation functions
11090 (BUF_strlcpy() and BUF_strlcat()).
11092 *Ben Laurie (CHATS) and Richard Levitte*
11094 * Avoid using fixed-size buffers for one-line DNs.
11096 *Ben Laurie (CHATS)*
11098 * Add BUF_MEM_grow_clean() to avoid information leakage when
11099 resizing buffers containing secrets, and use where appropriate.
11101 *Ben Laurie (CHATS)*
11103 * Avoid using fixed size buffers for configuration file location.
11105 *Ben Laurie (CHATS)*
11107 * Avoid filename truncation for various CA files.
11109 *Ben Laurie (CHATS)*
11111 * Use sizeof in preference to magic numbers.
11113 *Ben Laurie (CHATS)*
11115 * Avoid filename truncation in cert requests.
11117 *Ben Laurie (CHATS)*
11119 * Add assertions to check for (supposedly impossible) buffer
11122 *Ben Laurie (CHATS)*
11124 * Don't cache truncated DNS entries in the local cache (this could
11125 potentially lead to a spoofing attack).
11127 *Ben Laurie (CHATS)*
11129 * Fix various buffers to be large enough for hex/decimal
11130 representations in a platform independent manner.
11132 *Ben Laurie (CHATS)*
11134 * Add CRYPTO_realloc_clean() to avoid information leakage when
11135 resizing buffers containing secrets, and use where appropriate.
11137 *Ben Laurie (CHATS)*
11139 * Add BIO_indent() to avoid much slightly worrying code to do
11142 *Ben Laurie (CHATS)*
11144 * Convert sprintf()/BIO_puts() to BIO_printf().
11146 *Ben Laurie (CHATS)*
11148 * buffer_gets() could terminate with the buffer only half
11151 *Ben Laurie (CHATS)*
11153 * Add assertions to prevent user-supplied crypto functions from
11154 overflowing internal buffers by having large block sizes, etc.
11156 *Ben Laurie (CHATS)*
11158 * New OPENSSL_assert() macro (similar to assert(), but enabled
11161 *Ben Laurie (CHATS)*
11163 * Eliminate unused copy of key in RC4.
11165 *Ben Laurie (CHATS)*
11167 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11169 *Ben Laurie (CHATS)*
11171 * Fix off-by-one error in EGD path.
11173 *Ben Laurie (CHATS)*
11175 * If RANDFILE path is too long, ignore instead of truncating.
11177 *Ben Laurie (CHATS)*
11179 * Eliminate unused and incorrectly sized X.509 structure
11182 *Ben Laurie (CHATS)*
11184 * Eliminate unused and dangerous function knumber().
11186 *Ben Laurie (CHATS)*
11188 * Eliminate unused and dangerous structure, KSSL_ERR.
11190 *Ben Laurie (CHATS)*
11192 * Protect against overlong session ID context length in an encoded
11193 session object. Since these are local, this does not appear to be
11196 *Ben Laurie (CHATS)*
11198 * Change from security patch (see 0.9.6e below) that did not affect
11199 the 0.9.6 release series:
11201 Remote buffer overflow in SSL3 protocol - an attacker could
11202 supply an oversized master key in Kerberos-enabled versions.
11205 *Ben Laurie (CHATS)*
11207 * Change the SSL kerb5 codes to match RFC 2712.
11211 * Make -nameopt work fully for req and add -reqopt switch.
11213 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11215 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11217 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11219 * Make sure tests can be performed even if the corresponding algorithms
11220 have been removed entirely. This was also the last step to make
11221 OpenSSL compilable with DJGPP under all reasonable conditions.
11223 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11225 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11226 to allow version independent disabling of normally unselected ciphers,
11227 which may be activated as a side-effect of selecting a single cipher.
11229 (E.g., cipher list string "RSA" enables ciphersuites that are left
11230 out of "ALL" because they do not provide symmetric encryption.
11231 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11233 *Lutz Jaenicke, Bodo Moeller*
11235 * Add appropriate support for separate platform-dependent build
11236 directories. The recommended way to make a platform-dependent
11237 build directory is the following (tested on Linux), maybe with
11240 # Place yourself outside of the OpenSSL source tree. In
11241 # this example, the environment variable OPENSSL_SOURCE
11242 # is assumed to contain the absolute OpenSSL source directory.
11243 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11244 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11245 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11246 mkdir -p `dirname $F`
11247 ln -s $OPENSSL_SOURCE/$F $F
11250 To be absolutely sure not to disturb the source tree, a "make clean"
11251 is a good thing. If it isn't successful, don't worry about it,
11252 it probably means the source directory is very clean.
11256 * Make sure any ENGINE control commands make local copies of string
11257 pointers passed to them whenever necessary. Otherwise it is possible
11258 the caller may have overwritten (or deallocated) the original string
11259 data when a later ENGINE operation tries to use the stored values.
11261 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11263 * Improve diagnostics in file reading and command-line digests.
11265 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11267 * Add AES modes CFB and OFB to the object database. Correct an
11268 error in AES-CFB decryption.
11272 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11273 allows existing EVP_CIPHER_CTX structures to be reused after
11274 calling `EVP_*Final()`. This behaviour is used by encryption
11275 BIOs and some applications. This has the side effect that
11276 applications must explicitly clean up cipher contexts with
11277 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11281 * Check the values of dna and dnb in bn_mul_recursive before calling
11282 bn_mul_comba (a non zero value means the a or b arrays do not contain
11283 n2 elements) and fallback to bn_mul_normal if either is not zero.
11287 * Fix escaping of non-ASCII characters when using the -subj option
11288 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11292 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11293 form for "surname", serialNumber has no short form.
11294 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11295 therefore remove "mail" short name for "internet 7".
11296 The OID for unique identifiers in X509 certificates is
11297 x500UniqueIdentifier, not uniqueIdentifier.
11298 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11302 * Add an "init" command to the ENGINE config module and auto initialize
11303 ENGINEs. Without any "init" command the ENGINE will be initialized
11304 after all ctrl commands have been executed on it. If init=1 the
11305 ENGINE is initialized at that point (ctrls before that point are run
11306 on the uninitialized ENGINE and after on the initialized one). If
11307 init=0 then the ENGINE will not be initialized at all.
11311 * Fix the 'app_verify_callback' interface so that the user-defined
11312 argument is actually passed to the callback: In the
11313 SSL_CTX_set_cert_verify_callback() prototype, the callback
11314 declaration has been changed from
11317 int (*cb)(X509_STORE_CTX *,void *);
11318 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11319 i=s->ctx->app_verify_callback(&ctx)
11320 has been changed into
11321 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11323 To update applications using SSL_CTX_set_cert_verify_callback(),
11324 a dummy argument can be added to their callback functions.
11326 *D. K. Smetters <smetters@parc.xerox.com>*
11328 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11330 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11332 * Add and OPENSSL_LOAD_CONF define which will cause
11333 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11334 This allows older applications to transparently support certain
11335 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11336 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11337 load the config file and OPENSSL_add_all_algorithms_conf() which will
11338 always load it have also been added.
11342 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11343 Adjust NIDs and EVP layer.
11345 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11347 * Config modules support in openssl utility.
11349 Most commands now load modules from the config file,
11350 though in a few (such as version) this isn't done
11351 because it couldn't be used for anything.
11353 In the case of ca and req the config file used is
11354 the same as the utility itself: that is the -config
11355 command line option can be used to specify an
11360 * Move default behaviour from OPENSSL_config(). If appname is NULL
11361 use "openssl_conf" if filename is NULL use default openssl config file.
11365 * Add an argument to OPENSSL_config() to allow the use of an alternative
11366 config section name. Add a new flag to tolerate a missing config file
11367 and move code to CONF_modules_load_file().
11371 * Support for crypto accelerator cards from Accelerated Encryption
11372 Processing, www.aep.ie. (Use engine 'aep')
11373 The support was copied from 0.9.6c [engine] and adapted/corrected
11374 to work with the new engine framework.
11376 *AEP Inc. and Richard Levitte*
11378 * Support for SureWare crypto accelerator cards from Baltimore
11379 Technologies. (Use engine 'sureware')
11380 The support was copied from 0.9.6c [engine] and adapted
11381 to work with the new engine framework.
11385 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11386 make the newer ENGINE framework commands for the CHIL engine work.
11388 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11390 * Make it possible to produce shared libraries on ReliantUNIX.
11392 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11394 * Add the configuration target debug-linux-ppro.
11395 Make 'openssl rsa' use the general key loading routines
11396 implemented in apps.c, and make those routines able to
11397 handle the key format FORMAT_NETSCAPE and the variant
11400 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11402 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11404 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11406 * Add -keyform to rsautl, and document -engine.
11408 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11410 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11411 BIO_R_NO_SUCH_FILE error code rather than the generic
11412 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11416 * Add new functions
11417 ERR_peek_last_error
11418 ERR_peek_last_error_line
11419 ERR_peek_last_error_line_data.
11420 These are similar to
11422 ERR_peek_error_line
11423 ERR_peek_error_line_data,
11424 but report on the latest error recorded rather than the first one
11425 still in the error queue.
11427 *Ben Laurie, Bodo Moeller*
11429 * default_algorithms option in ENGINE config module. This allows things
11431 default_algorithms = ALL
11432 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11436 * Preliminary ENGINE config module.
11440 * New experimental application configuration code.
11444 * Change the AES code to follow the same name structure as all other
11445 symmetric ciphers, and behave the same way. Move everything to
11446 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11448 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11450 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11452 *Ben Laurie and Theo de Raadt*
11454 * Add option to output public keys in req command.
11456 *Massimiliano Pala madwolf@openca.org*
11458 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11459 (up to about 10% better than before for P-192 and P-224).
11463 * New functions/macros
11465 SSL_CTX_set_msg_callback(ctx, cb)
11466 SSL_CTX_set_msg_callback_arg(ctx, arg)
11467 SSL_set_msg_callback(ssl, cb)
11468 SSL_set_msg_callback_arg(ssl, arg)
11470 to request calling a callback function
11472 void cb(int write_p, int version, int content_type,
11473 const void *buf, size_t len, SSL *ssl, void *arg)
11475 whenever a protocol message has been completely received
11476 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11477 protocol version according to which the SSL library interprets
11478 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11479 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11480 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11481 specification (change_cipher_spec(20), alert(21), handshake(22)).
11482 'buf' and 'len' point to the actual message, 'ssl' to the
11483 SSL object, and 'arg' is the application-defined value set by
11484 SSL[_CTX]_set_msg_callback_arg().
11486 'openssl s_client' and 'openssl s_server' have new '-msg' options
11487 to enable a callback that displays all protocol messages.
11491 * Change the shared library support so shared libraries are built as
11492 soon as the corresponding static library is finished, and thereby get
11493 openssl and the test programs linked against the shared library.
11494 This still only happens when the keyword "shard" has been given to
11495 the configuration scripts.
11497 NOTE: shared library support is still an experimental thing, and
11498 backward binary compatibility is still not guaranteed.
11500 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11502 * Add support for Subject Information Access extension.
11504 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11506 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11507 additional bytes when new memory had to be allocated, not just
11508 when reusing an existing buffer.
11512 * New command line and configuration option 'utf8' for the req command.
11513 This allows field values to be specified as UTF8 strings.
11517 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11518 runs for the former and machine-readable output for the latter.
11522 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11523 of the e-mail address in the DN (i.e., it will go into a certificate
11524 extension only). The new configuration file option 'email_in_dn = no'
11525 has the same effect.
11527 *Massimiliano Pala madwolf@openca.org*
11529 * Change all functions with names starting with `des_` to be starting
11530 with `DES_` instead. Add wrappers that are compatible with libdes,
11531 but are named `_ossl_old_des_*`. Finally, add macros that map the
11532 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11533 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11534 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11537 Since we provide two compatibility mappings, the user needs to
11538 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11539 compatibility is desired. The default (i.e., when that macro
11540 isn't defined) is OpenSSL 0.9.6c compatibility.
11542 There are also macros that enable and disable the support of old
11543 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11544 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11545 are defined, the default will apply: to support the old des routines.
11547 In either case, one must include openssl/des.h to get the correct
11548 definitions. Do not try to just include openssl/des_old.h, that
11551 NOTE: This is a major break of an old API into a new one. Software
11552 authors are encouraged to switch to the `DES_` style functions. Some
11553 time in the future, des_old.h and the libdes compatibility functions
11554 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11555 default), and then completely removed.
11559 * Test for certificates which contain unsupported critical extensions.
11560 If such a certificate is found during a verify operation it is
11561 rejected by default: this behaviour can be overridden by either
11562 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11563 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11564 X509_supported_extension() has also been added which returns 1 if a
11565 particular extension is supported.
11569 * Modify the behaviour of EVP cipher functions in similar way to digests
11570 to retain compatibility with existing code.
11574 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11575 compatibility with existing code. In particular the 'ctx' parameter does
11576 not have to be to be initialized before the call to EVP_DigestInit() and
11577 it is tidied up after a call to EVP_DigestFinal(). New function
11578 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11579 EVP_MD_CTX_copy() changed to not require the destination to be
11580 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11581 requires the destination to be valid.
11583 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11584 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11588 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11589 so that complete 'Handshake' protocol structures are kept in memory
11590 instead of overwriting 'msg_type' and 'length' with 'body' data.
11594 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11596 *Massimo Santin via Richard Levitte*
11598 * Major restructuring to the underlying ENGINE code. This includes
11599 reduction of linker bloat, separation of pure "ENGINE" manipulation
11600 (initialisation, etc) from functionality dealing with implementations
11601 of specific crypto interfaces. This change also introduces integrated
11602 support for symmetric ciphers and digest implementations - so ENGINEs
11603 can now accelerate these by providing EVP_CIPHER and EVP_MD
11604 implementations of their own. This is detailed in
11605 [crypto/engine/README.md](crypto/engine/README.md)
11606 as it couldn't be adequately described here. However, there are a few
11607 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11608 were changed in the original introduction of ENGINE code have now
11609 reverted back - the hooking from this code to ENGINE is now a good
11610 deal more passive and at run-time, operations deal directly with
11611 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11612 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11613 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11614 they were not being used by the framework as there is no concept of a
11615 BIGNUM_METHOD and they could not be generalised to the new
11616 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11617 ENGINE_cpy() has been removed as it cannot be consistently defined in
11622 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11626 * Change mkdef.pl to sort symbols that get the same entry number,
11627 and make sure the automatically generated functions `ERR_load_*`
11628 become part of libeay.num as well.
11632 * New function SSL_renegotiate_pending(). This returns true once
11633 renegotiation has been requested (either SSL_renegotiate() call
11634 or HelloRequest/ClientHello received from the peer) and becomes
11635 false once a handshake has been completed.
11636 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11637 sends a HelloRequest, but does not ensure that a handshake takes
11638 place. SSL_renegotiate_pending() is useful for checking if the
11639 client has followed the request.)
11643 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11644 By default, clients may request session resumption even during
11645 renegotiation (if session ID contexts permit); with this option,
11646 session resumption is possible only in the first handshake.
11648 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11649 more bits available for options that should not be part of
11650 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11654 * Add some demos for certificate and certificate request creation.
11658 * Make maximum certificate chain size accepted from the peer application
11659 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11660 "Douglas E. Engert" <deengert@anl.gov>.
11664 * Add support for shared libraries for Unixware-7
11665 (Boyd Lynn Gerber <gerberb@zenez.com>).
11669 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11670 be done prior to destruction. Use this to unload error strings from
11671 ENGINEs that load their own error strings. NB: This adds two new API
11672 functions to "get" and "set" this destroy handler in an ENGINE.
11676 * Alter all existing ENGINE implementations (except "openssl" and
11677 "openbsd") to dynamically instantiate their own error strings. This
11678 makes them more flexible to be built both as statically-linked ENGINEs
11679 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11680 Also, add stub code to each that makes building them as self-contained
11681 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11685 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11686 implementations into applications that are completely implemented in
11687 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11688 commands that can be used to configure what shared-library to load and
11689 to control aspects of the way it is handled. Also, made an update to
11690 the [README-Engine.md](README-Engine.md) file
11691 that brings its information up-to-date and
11692 provides some information and instructions on the "dynamic" ENGINE
11693 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11697 * Make it possible to unload ranges of ERR strings with a new
11698 "ERR_unload_strings" function.
11702 * Add a copy() function to EVP_MD.
11706 * Make EVP_MD routines take a context pointer instead of just the
11707 md_data void pointer.
11711 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11712 that the digest can only process a single chunk of data
11713 (typically because it is provided by a piece of
11714 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11715 is only going to provide a single chunk of data, and hence the
11716 framework needn't accumulate the data for oneshot drivers.
11720 * As with "ERR", make it possible to replace the underlying "ex_data"
11721 functions. This change also alters the storage and management of global
11722 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11723 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11724 index counters. The API functions that use this state have been changed
11725 to take a "class_index" rather than pointers to the class's local STACK
11726 and counter, and there is now an API function to dynamically create new
11727 classes. This centralisation allows us to (a) plug a lot of the
11728 thread-safety problems that existed, and (b) makes it possible to clean
11729 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11730 such data would previously have always leaked in application code and
11731 workarounds were in place to make the memory debugging turn a blind eye
11732 to it. Application code that doesn't use this new function will still
11733 leak as before, but their memory debugging output will announce it now
11734 rather than letting it slide.
11736 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11737 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11738 has a return value to indicate success or failure.
11742 * Make it possible to replace the underlying "ERR" functions such that the
11743 global state (2 LHASH tables and 2 locks) is only used by the "default"
11744 implementation. This change also adds two functions to "get" and "set"
11745 the implementation prior to it being automatically set the first time
11746 any other ERR function takes place. Ie. an application can call "get",
11747 pass the return value to a module it has just loaded, and that module
11748 can call its own "set" function using that value. This means the
11749 module's "ERR" operations will use (and modify) the error state in the
11750 application and not in its own statically linked copy of OpenSSL code.
11754 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11755 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11756 the operation, and provides a more encapsulated way for external code
11757 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11758 to use these functions rather than manually incrementing the counts.
11760 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11764 * Add EVP test program.
11768 * Add symmetric cipher support to ENGINE. Expect the API to change!
11772 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11773 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11774 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11775 These allow a CRL to be built without having to access X509_CRL fields
11776 directly. Modify 'ca' application to use new functions.
11780 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11781 bug workarounds. Rollback attack detection is a security feature.
11782 The problem will only arise on OpenSSL servers when TLSv1 is not
11783 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11784 Software authors not wanting to support TLSv1 will have special reasons
11785 for their choice and can explicitly enable this option.
11787 *Bodo Moeller, Lutz Jaenicke*
11789 * Rationalise EVP so it can be extended: don't include a union of
11790 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11791 (similar to those existing for EVP_CIPHER_CTX).
11796 EVP_MD_CTX_init(&md); /* new function call */
11797 EVP_DigestInit(&md, EVP_sha1());
11798 EVP_DigestUpdate(&md, in, len);
11799 EVP_DigestFinal(&md, out, NULL);
11800 EVP_MD_CTX_cleanup(&md); /* new function call */
11804 * Make DES key schedule conform to the usual scheme, as well as
11805 correcting its structure. This means that calls to DES functions
11806 now have to pass a pointer to a des_key_schedule instead of a
11807 plain des_key_schedule (which was actually always a pointer
11810 des_key_schedule ks;
11812 des_set_key_checked(..., &ks);
11813 des_ncbc_encrypt(..., &ks, ...);
11815 (Note that a later change renames 'des_...' into 'DES_...'.)
11819 * Initial reduction of linker bloat: the use of some functions, such as
11820 PEM causes large amounts of unused functions to be linked in due to
11821 poor organisation. For example pem_all.c contains every PEM function
11822 which has a knock on effect of linking in large amounts of (unused)
11823 ASN1 code. Grouping together similar functions and splitting unrelated
11824 functions prevents this.
11828 * Cleanup of EVP macros.
11832 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
11833 correct `_ecb suffix`.
11837 * Add initial OCSP responder support to ocsp application. The
11838 revocation information is handled using the text based index
11839 use by the ca application. The responder can either handle
11840 requests generated internally, supplied in files (for example
11841 via a CGI script) or using an internal minimal server.
11845 * Add configuration choices to get zlib compression for TLS.
11849 * Changes to Kerberos SSL for RFC 2712 compliance:
11850 1. Implemented real KerberosWrapper, instead of just using
11851 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
11852 2. Implemented optional authenticator field of KerberosWrapper.
11854 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
11855 and authenticator structs; see crypto/krb5/.
11857 Generalized Kerberos calls to support multiple Kerberos libraries.
11858 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
11859 via Richard Levitte*
11861 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
11862 already does with RSA. testdsa.h now has 'priv_key/pub_key'
11863 values for each of the key sizes rather than having just
11864 parameters (and 'speed' generating keys each time).
11868 * Speed up EVP routines.
11871 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
11872 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
11873 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
11874 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
11876 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
11877 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
11878 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
11881 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
11883 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
11887 * Added the OS2-EMX target.
11889 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
11891 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
11892 to support NCONF routines in extension code. New function CONF_set_nconf()
11893 to allow functions which take an NCONF to also handle the old LHASH
11894 structure: this means that the old CONF compatible routines can be
11895 retained (in particular wrt extensions) without having to duplicate the
11896 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
11900 * Enhance the general user interface with mechanisms for inner control
11901 and with possibilities to have yes/no kind of prompts.
11905 * Change all calls to low level digest routines in the library and
11906 applications to use EVP. Add missing calls to HMAC_cleanup() and
11907 don't assume HMAC_CTX can be copied using memcpy().
11909 *Verdon Walker <VWalker@novell.com>, Steve Henson*
11911 * Add the possibility to control engines through control names but with
11912 arbitrary arguments instead of just a string.
11913 Change the key loaders to take a UI_METHOD instead of a callback
11914 function pointer. NOTE: this breaks binary compatibility with earlier
11915 versions of OpenSSL [engine].
11916 Adapt the nCipher code for these new conditions and add a card insertion
11921 * Enhance the general user interface with mechanisms to better support
11922 dialog box interfaces, application-defined prompts, the possibility
11923 to use defaults (for example default passwords from somewhere else)
11924 and interrupts/cancellations.
11928 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
11929 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
11933 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
11934 tidy up some unnecessarily weird code in 'sk_new()').
11936 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
11938 * Change the key loading routines for ENGINEs to use the same kind
11939 callback (pem_password_cb) as all other routines that need this
11944 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
11945 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
11946 than this minimum value is recommended.
11950 * New random seeder for OpenVMS, using the system process statistics
11951 that are easily reachable.
11955 * Windows apparently can't transparently handle global
11956 variables defined in DLLs. Initialisations such as:
11958 const ASN1_ITEM *it = &ASN1_INTEGER_it;
11960 won't compile. This is used by the any applications that need to
11961 declare their own ASN1 modules. This was fixed by adding the option
11962 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
11963 needed for static libraries under Win32.
11967 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
11968 setting of purpose and trust fields. New X509_STORE trust and
11969 purpose functions and tidy up setting in other SSL functions.
11973 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
11974 structure. These are inherited by X509_STORE_CTX when it is
11975 initialised. This allows various defaults to be set in the
11976 X509_STORE structure (such as flags for CRL checking and custom
11977 purpose or trust settings) for functions which only use X509_STORE_CTX
11978 internally such as S/MIME.
11980 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
11981 trust settings if they are not set in X509_STORE. This allows X509_STORE
11982 purposes and trust (in S/MIME for example) to override any set by default.
11984 Add command line options for CRL checking to smime, s_client and s_server
11989 * Initial CRL based revocation checking. If the CRL checking flag(s)
11990 are set then the CRL is looked up in the X509_STORE structure and
11991 its validity and signature checked, then if the certificate is found
11992 in the CRL the verify fails with a revoked error.
11994 Various new CRL related callbacks added to X509_STORE_CTX structure.
11996 Command line options added to 'verify' application to support this.
11998 This needs some additional work, such as being able to handle multiple
11999 CRLs with different times, extension based lookup (rather than just
12000 by subject name) and ultimately more complete V2 CRL extension
12005 * Add a general user interface API (crypto/ui/). This is designed
12006 to replace things like des_read_password and friends (backward
12007 compatibility functions using this new API are provided).
12008 The purpose is to remove prompting functions from the DES code
12009 section as well as provide for prompting through dialog boxes in
12010 a window system and the like.
12014 * Add "ex_data" support to ENGINE so implementations can add state at a
12015 per-structure level rather than having to store it globally.
12019 * Make it possible for ENGINE structures to be copied when retrieved by
12020 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12021 This causes the "original" ENGINE structure to act like a template,
12022 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12023 operational state can be localised to each ENGINE structure, despite the
12024 fact they all share the same "methods". New ENGINE structures returned in
12025 this case have no functional references and the return value is the single
12026 structural reference. This matches the single structural reference returned
12027 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12032 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12033 needs to match any other type at all we need to manually clear the
12038 * Changes to the "openssl engine" utility to include;
12039 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12040 about an ENGINE's available control commands.
12041 - executing control commands from command line arguments using the
12042 '-pre' and '-post' switches. '-post' is only used if '-t' is
12043 specified and the ENGINE is successfully initialised. The syntax for
12044 the individual commands are colon-separated, for example;
12045 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12049 * New dynamic control command support for ENGINEs. ENGINEs can now
12050 declare their own commands (numbers), names (strings), descriptions,
12051 and input types for run-time discovery by calling applications. A
12052 subset of these commands are implicitly classed as "executable"
12053 depending on their input type, and only these can be invoked through
12054 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12055 can be based on user input, config files, etc). The distinction is
12056 that "executable" commands cannot return anything other than a boolean
12057 result and can only support numeric or string input, whereas some
12058 discoverable commands may only be for direct use through
12059 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12060 pointers, or other custom uses. The "executable" commands are to
12061 support parameterisations of ENGINE behaviour that can be
12062 unambiguously defined by ENGINEs and used consistently across any
12063 OpenSSL-based application. Commands have been added to all the
12064 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12065 control over shared-library paths without source code alterations.
12069 * Changed all ENGINE implementations to dynamically allocate their
12070 ENGINEs rather than declaring them statically. Apart from this being
12071 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12072 this also allows the implementations to compile without using the
12073 internal engine_int.h header.
12077 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12078 'const' value. Any code that should be able to modify a RAND_METHOD
12079 should already have non-const pointers to it (ie. they should only
12080 modify their own ones).
12084 * Made a variety of little tweaks to the ENGINE code.
12085 - "atalla" and "ubsec" string definitions were moved from header files
12086 to C code. "nuron" string definitions were placed in variables
12087 rather than hard-coded - allowing parameterisation of these values
12088 later on via ctrl() commands.
12089 - Removed unused "#if 0"'d code.
12090 - Fixed engine list iteration code so it uses ENGINE_free() to release
12091 structural references.
12092 - Constified the RAND_METHOD element of ENGINE structures.
12093 - Constified various get/set functions as appropriate and added
12094 missing functions (including a catch-all ENGINE_cpy that duplicates
12095 all ENGINE values onto a new ENGINE except reference counts/state).
12096 - Removed NULL parameter checks in get/set functions. Setting a method
12097 or function to NULL is a way of cancelling out a previously set
12098 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12099 and doesn't justify the extra error symbols and code.
12100 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12101 flags from engine_int.h to engine.h.
12102 - Changed prototypes for ENGINE handler functions (init(), finish(),
12103 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12107 * Implement binary inversion algorithm for BN_mod_inverse in addition
12108 to the algorithm using long division. The binary algorithm can be
12109 used only if the modulus is odd. On 32-bit systems, it is faster
12110 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12111 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12112 up to 450 bits. In 64-bit environments, the binary algorithm
12113 appears to be advantageous for much longer moduli; here we use it
12114 for moduli up to 2048 bits.
12118 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12119 could not support the combine flag in choice fields.
12123 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12124 extensions from a certificate request to the certificate.
12128 * Allow multiple 'certopt' and 'nameopt' options to be separated
12129 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12130 file: this allows the display of the certificate about to be
12131 signed to be customised, to allow certain fields to be included
12132 or excluded and extension details. The old system didn't display
12133 multicharacter strings properly, omitted fields not in the policy
12134 and couldn't display additional details such as extensions.
12138 * Function EC_POINTs_mul for multiple scalar multiplication
12139 of an arbitrary number of elliptic curve points
12140 \sum scalars[i]*points[i],
12141 optionally including the generator defined for the EC_GROUP:
12142 scalar*generator + \sum scalars[i]*points[i].
12144 EC_POINT_mul is a simple wrapper function for the typical case
12145 that the point list has just one item (besides the optional
12150 * First EC_METHODs for curves over GF(p):
12152 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12153 operations and provides various method functions that can also
12154 operate with faster implementations of modular arithmetic.
12156 EC_GFp_mont_method() reuses most functions that are part of
12157 EC_GFp_simple_method, but uses Montgomery arithmetic.
12159 *Bodo Moeller; point addition and point doubling
12160 implementation directly derived from source code provided by
12161 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12163 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12164 crypto/ec/ec_lib.c):
12166 Curves are EC_GROUP objects (with an optional group generator)
12167 based on EC_METHODs that are built into the library.
12169 Points are EC_POINT objects based on EC_GROUP objects.
12171 Most of the framework would be able to handle curves over arbitrary
12172 finite fields, but as there are no obvious types for fields other
12173 than GF(p), some functions are limited to that for now.
12177 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12178 that the file contains a complete HTTP response.
12182 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12183 change the def and num file printf format specifier from "%-40sXXX"
12184 to "%-39s XXX". The latter will always guarantee a space after the
12185 field while the former will cause them to run together if the field
12186 is 40 of more characters long.
12190 * Constify the cipher and digest 'method' functions and structures
12191 and modify related functions to take constant EVP_MD and EVP_CIPHER
12196 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12197 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12201 * Modify `EVP_Digest*()` routines so they now return values. Although the
12202 internal software routines can never fail additional hardware versions
12207 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12209 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12210 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12213 ERR_R_NESTED_ASN1_ERROR
12215 ERR_R_MISSING_ASN1_EOS
12216 were 4 .. 9, conflicting with
12217 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12219 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12220 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12222 Add new error code 'ERR_R_INTERNAL_ERROR'.
12226 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12231 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12232 sets the subject name for a new request or supersedes the
12233 subject name in a given request. Formats that can be parsed are
12234 'CN=Some Name, OU=myOU, C=IT'
12236 'CN=Some Name/OU=myOU/C=IT'.
12238 Add options '-batch' and '-verbose' to 'openssl req'.
12240 *Massimiliano Pala <madwolf@hackmasters.net>*
12242 * Introduce the possibility to access global variables through
12243 functions on platform were that's the best way to handle exporting
12244 global variables in shared libraries. To enable this functionality,
12245 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12246 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12247 is normally done by Configure or something similar).
12249 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12250 in the source file (foo.c) like this:
12252 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12253 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12255 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12256 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12258 OPENSSL_DECLARE_GLOBAL(int,foo);
12259 #define foo OPENSSL_GLOBAL_REF(foo)
12260 OPENSSL_DECLARE_GLOBAL(double,bar);
12261 #define bar OPENSSL_GLOBAL_REF(bar)
12263 The #defines are very important, and therefore so is including the
12264 header file everywhere where the defined globals are used.
12266 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12267 of ASN.1 items, but that structure is a bit different.
12269 The largest change is in util/mkdef.pl which has been enhanced with
12270 better and easier to understand logic to choose which symbols should
12271 go into the Windows .def files as well as a number of fixes and code
12272 cleanup (among others, algorithm keywords are now sorted
12273 lexicographically to avoid constant rewrites).
12277 * In BN_div() keep a copy of the sign of 'num' before writing the
12278 result to 'rm' because if rm==num the value will be overwritten
12279 and produce the wrong result if 'num' is negative: this caused
12280 problems with BN_mod() and BN_nnmod().
12284 * Function OCSP_request_verify(). This checks the signature on an
12285 OCSP request and verifies the signer certificate. The signer
12286 certificate is just checked for a generic purpose and OCSP request
12291 * Add OCSP_check_validity() function to check the validity of OCSP
12292 responses. OCSP responses are prepared in real time and may only
12293 be a few seconds old. Simply checking that the current time lies
12294 between thisUpdate and nextUpdate max reject otherwise valid responses
12295 caused by either OCSP responder or client clock inaccuracy. Instead
12296 we allow thisUpdate and nextUpdate to fall within a certain period of
12297 the current time. The age of the response can also optionally be
12298 checked. Two new options -validity_period and -status_age added to
12303 * If signature or public key algorithm is unrecognized print out its
12304 OID rather that just UNKNOWN.
12308 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12309 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12310 ID to be generated from the issuer certificate alone which can then be
12311 passed to OCSP_id_issuer_cmp().
12315 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12316 ASN1 modules to export functions returning ASN1_ITEM pointers
12317 instead of the ASN1_ITEM structures themselves. This adds several
12318 new macros which allow the underlying ASN1 function/structure to
12319 be accessed transparently. As a result code should not use ASN1_ITEM
12320 references directly (such as &X509_it) but instead use the relevant
12321 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12322 use of the new ASN1 code on platforms where exporting structures
12323 is problematical (for example in shared libraries) but exporting
12324 functions returning pointers to structures is not.
12328 * Add support for overriding the generation of SSL/TLS session IDs.
12329 These callbacks can be registered either in an SSL_CTX or per SSL.
12330 The purpose of this is to allow applications to control, if they wish,
12331 the arbitrary values chosen for use as session IDs, particularly as it
12332 can be useful for session caching in multiple-server environments. A
12333 command-line switch for testing this (and any client code that wishes
12334 to use such a feature) has been added to "s_server".
12336 *Geoff Thorpe, Lutz Jaenicke*
12338 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12339 of the form `#if defined(...) || defined(...) || ...` and
12340 `#if !defined(...) && !defined(...) && ...`. This also avoids
12341 the growing number of special cases it was previously handling.
12345 * Make all configuration macros available for application by making
12346 sure they are available in opensslconf.h, by giving them names starting
12347 with `OPENSSL_` to avoid conflicts with other packages and by making
12348 sure e_os2.h will cover all platform-specific cases together with
12350 Additionally, it is now possible to define configuration/platform-
12351 specific names (called "system identities"). In the C code, these
12352 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12353 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12354 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12359 * New option -set_serial to 'req' and 'x509' this allows the serial
12360 number to use to be specified on the command line. Previously self
12361 signed certificates were hard coded with serial number 0 and the
12362 CA options of 'x509' had to use a serial number in a file which was
12367 * New options to 'ca' utility to support V2 CRL entry extensions.
12368 Currently CRL reason, invalidity date and hold instruction are
12369 supported. Add new CRL extensions to V3 code and some new objects.
12373 * New function EVP_CIPHER_CTX_set_padding() this is used to
12374 disable standard block padding (aka PKCS#5 padding) in the EVP
12375 API, which was previously mandatory. This means that the data is
12376 not padded in any way and so the total length much be a multiple
12377 of the block size, otherwise an error occurs.
12381 * Initial (incomplete) OCSP SSL support.
12385 * New function OCSP_parse_url(). This splits up a URL into its host,
12386 port and path components: primarily to parse OCSP URLs. New -url
12387 option to ocsp utility.
12391 * New nonce behavior. The return value of OCSP_check_nonce() now
12392 reflects the various checks performed. Applications can decide
12393 whether to tolerate certain situations such as an absent nonce
12394 in a response when one was present in a request: the ocsp application
12395 just prints out a warning. New function OCSP_add1_basic_nonce()
12396 this is to allow responders to include a nonce in a response even if
12397 the request is nonce-less.
12401 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12402 skipped when using openssl x509 multiple times on a single input file,
12403 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12407 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12408 set string type: to handle setting ASN1_TIME structures. Fix ca
12409 utility to correctly initialize revocation date of CRLs.
12413 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12414 the clients preferred ciphersuites and rather use its own preferences.
12415 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12416 Internet Explorer by ensuring unchanged hash method during stepup.
12417 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12421 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12422 to aes and add a new 'exist' option to print out symbols that don't
12427 * Additional options to ocsp utility to allow flags to be set and
12428 additional certificates supplied.
12432 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12433 OCSP client a number of certificate to only verify the response
12438 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12439 handle the new API. Currently only ECB, CBC modes supported. Add new
12442 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12443 Encryption Standard (AES) Ciphersuites for Transport Layer
12444 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12445 not enabled by default and were not part of the "ALL" ciphersuite
12446 alias because they were not yet official; they could be
12447 explicitly requested by specifying the "AESdraft" ciphersuite
12448 group alias. In the final release of OpenSSL 0.9.7, the group
12449 alias is called "AES" and is part of "ALL".)
12451 *Ben Laurie, Steve Henson, Bodo Moeller*
12453 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12454 request to response.
12458 * Functions for OCSP responders. OCSP_request_onereq_count(),
12459 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12460 extract information from a certificate request. OCSP_response_create()
12461 creates a response and optionally adds a basic response structure.
12462 OCSP_basic_add1_status() adds a complete single response to a basic
12463 response and returns the OCSP_SINGLERESP structure just added (to allow
12464 extensions to be included for example). OCSP_basic_add1_cert() adds a
12465 certificate to a basic response and OCSP_basic_sign() signs a basic
12466 response with various flags. New helper functions ASN1_TIME_check()
12467 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12468 (converts ASN1_TIME to GeneralizedTime).
12472 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12473 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12474 structure from a certificate. X509_pubkey_digest() digests the public_key
12475 contents: this is used in various key identifiers.
12479 * Make sk_sort() tolerate a NULL argument.
12481 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12483 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12484 passed by the function are trusted implicitly. If any of them signed the
12485 response then it is assumed to be valid and is not verified.
12489 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12490 to data. This was previously part of the PKCS7 ASN1 code. This
12491 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12492 *Steve Henson, reported by Kenneth R. Robinette
12493 <support@securenetterm.com>*
12495 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12496 routines: without these tracing memory leaks is very painful.
12497 Fix leaks in PKCS12 and PKCS7 routines.
12501 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12502 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12503 effectively meant GeneralizedTime would never be used. Now it
12504 is initialised to -1 but X509_time_adj() now has to check the value
12505 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12506 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12507 *Steve Henson, reported by Kenneth R. Robinette
12508 <support@securenetterm.com>*
12510 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12511 result in a zero length in the ASN1_INTEGER structure which was
12512 not consistent with the structure when d2i_ASN1_INTEGER() was used
12513 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12514 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12515 where it did not print out a minus for negative ASN1_INTEGER.
12519 * Add summary printout to ocsp utility. The various functions which
12520 convert status values to strings have been renamed to:
12521 OCSP_response_status_str(), OCSP_cert_status_str() and
12522 OCSP_crl_reason_str() and are no longer static. New options
12523 to verify nonce values and to disable verification. OCSP response
12524 printout format cleaned up.
12528 * Add additional OCSP certificate checks. These are those specified
12529 in RFC2560. This consists of two separate checks: the CA of the
12530 certificate being checked must either be the OCSP signer certificate
12531 or the issuer of the OCSP signer certificate. In the latter case the
12532 OCSP signer certificate must contain the OCSP signing extended key
12533 usage. This check is performed by attempting to match the OCSP
12534 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12535 in the OCSP_CERTID structures of the response.
12539 * Initial OCSP certificate verification added to OCSP_basic_verify()
12540 and related routines. This uses the standard OpenSSL certificate
12541 verify routines to perform initial checks (just CA validity) and
12542 to obtain the certificate chain. Then additional checks will be
12543 performed on the chain. Currently the root CA is checked to see
12544 if it is explicitly trusted for OCSP signing. This is used to set
12545 a root CA as a global signing root: that is any certificate that
12546 chains to that CA is an acceptable OCSP signing certificate.
12550 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12551 extensions from a separate configuration file.
12552 As when reading extensions from the main configuration file,
12553 the '-extensions ...' option may be used for specifying the
12556 *Massimiliano Pala <madwolf@comune.modena.it>*
12558 * New OCSP utility. Allows OCSP requests to be generated or
12559 read. The request can be sent to a responder and the output
12560 parsed, outputted or printed in text form. Not complete yet:
12561 still needs to check the OCSP response validity.
12565 * New subcommands for 'openssl ca':
12566 `openssl ca -status <serial>` prints the status of the cert with
12567 the given serial number (according to the index file).
12568 `openssl ca -updatedb` updates the expiry status of certificates
12571 *Massimiliano Pala <madwolf@comune.modena.it>*
12573 * New '-newreq-nodes' command option to CA.pl. This is like
12574 '-newreq', but calls 'openssl req' with the '-nodes' option
12575 so that the resulting key is not encrypted.
12577 *Damien Miller <djm@mindrot.org>*
12579 * New configuration for the GNU Hurd.
12581 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12583 * Initial code to implement OCSP basic response verify. This
12584 is currently incomplete. Currently just finds the signer's
12585 certificate and verifies the signature on the response.
12589 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12590 value of OPENSSLDIR. This is available via the new '-d' option
12591 to 'openssl version', and is also included in 'openssl version -a'.
12595 * Allowing defining memory allocation callbacks that will be given
12596 file name and line number information in additional arguments
12597 (a `const char*` and an int). The basic functionality remains, as
12598 well as the original possibility to just replace malloc(),
12599 realloc() and free() by functions that do not know about these
12600 additional arguments. To register and find out the current
12601 settings for extended allocation functions, the following
12602 functions are provided:
12604 CRYPTO_set_mem_ex_functions
12605 CRYPTO_set_locked_mem_ex_functions
12606 CRYPTO_get_mem_ex_functions
12607 CRYPTO_get_locked_mem_ex_functions
12609 These work the same way as CRYPTO_set_mem_functions and friends.
12610 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12611 extended allocation function is enabled.
12612 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12613 a conventional allocation function is enabled.
12615 *Richard Levitte, Bodo Moeller*
12617 * Finish off removing the remaining LHASH function pointer casts.
12618 There should no longer be any prototype-casting required when using
12619 the LHASH abstraction, and any casts that remain are "bugs". See
12620 the callback types and macros at the head of lhash.h for details
12621 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12625 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12626 If /dev/[u]random devices are not available or do not return enough
12627 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12629 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12630 /etc/entropy will be queried once each in this sequence, querying stops
12631 when enough entropy was collected without querying more sockets.
12635 * Change the Unix RAND_poll() variant to be able to poll several
12636 random devices, as specified by DEVRANDOM, until a sufficient amount
12637 of data has been collected. We spend at most 10 ms on each file
12638 (select timeout) and read in non-blocking mode. DEVRANDOM now
12639 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12640 (previously it was just the string "/dev/urandom"), so on typical
12641 platforms the 10 ms delay will never occur.
12642 Also separate out the Unix variant to its own file, rand_unix.c.
12643 For VMS, there's a currently-empty rand_vms.c.
12647 * Move OCSP client related routines to ocsp_cl.c. These
12648 provide utility functions which an application needing
12649 to issue a request to an OCSP responder and analyse the
12650 response will typically need: as opposed to those which an
12651 OCSP responder itself would need which will be added later.
12653 OCSP_request_sign() signs an OCSP request with an API similar
12654 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12655 response. OCSP_response_get1_basic() extracts basic response
12656 from response. OCSP_resp_find_status(): finds and extracts status
12657 information from an OCSP_CERTID structure (which will be created
12658 when the request structure is built). These are built from lower
12659 level functions which work on OCSP_SINGLERESP structures but
12660 won't normally be used unless the application wishes to examine
12661 extensions in the OCSP response for example.
12663 Replace nonce routines with a pair of functions.
12664 OCSP_request_add1_nonce() adds a nonce value and optionally
12665 generates a random value. OCSP_check_nonce() checks the
12666 validity of the nonce in an OCSP response.
12670 * Change function OCSP_request_add() to OCSP_request_add0_id().
12671 This doesn't copy the supplied OCSP_CERTID and avoids the
12672 need to free up the newly created id. Change return type
12673 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12674 This can then be used to add extensions to the request.
12675 Deleted OCSP_request_new(), since most of its functionality
12676 is now in OCSP_REQUEST_new() (and the case insensitive name
12677 clash) apart from the ability to set the request name which
12678 will be added elsewhere.
12682 * Update OCSP API. Remove obsolete extensions argument from
12683 various functions. Extensions are now handled using the new
12684 OCSP extension code. New simple OCSP HTTP function which
12685 can be used to send requests and parse the response.
12689 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12690 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12691 uses the special reorder version of SET OF to sort the attributes
12692 and reorder them to match the encoded order. This resolves a long
12693 standing problem: a verify on a PKCS7 structure just after signing
12694 it used to fail because the attribute order did not match the
12695 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12696 it uses the received order. This is necessary to tolerate some broken
12697 software that does not order SET OF. This is handled by encoding
12698 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12699 to produce the required SET OF.
12703 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12704 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12705 files to get correct declarations of the ASN.1 item variables.
12709 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12710 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12711 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12712 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12713 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12714 ASN1_ITEM and no wrapper functions.
12718 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12719 replace the old function pointer based I/O routines. Change most of
12720 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12724 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12725 lines, recognize more "algorithms" that can be deselected, and make
12726 it complain about algorithm deselection that isn't recognised.
12730 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12731 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12732 to use new functions. Add NO_ASN1_OLD which can be set to remove
12733 some old style ASN1 functions: this can be used to determine if old
12734 code will still work when these eventually go away.
12738 * New extension functions for OCSP structures, these follow the
12739 same conventions as certificates and CRLs.
12743 * New function X509V3_add1_i2d(). This automatically encodes and
12744 adds an extension. Its behaviour can be customised with various
12745 flags to append, replace or delete. Various wrappers added for
12746 certificates and CRLs.
12750 * Fix to avoid calling the underlying ASN1 print routine when
12751 an extension cannot be parsed. Correct a typo in the
12752 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12756 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12757 entries for variables.
12761 * Add functionality to apps/openssl.c for detecting locking
12762 problems: As the program is single-threaded, all we have
12763 to do is register a locking callback using an array for
12764 storing which locks are currently held by the program.
12768 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12769 SSL_get_ex_data_X509_STORE_idx(), which is used in
12770 ssl_verify_cert_chain() and thus can be called at any time
12771 during TLS/SSL handshakes so that thread-safety is essential.
12772 Unfortunately, the ex_data design is not at all suited
12773 for multi-threaded use, so it probably should be abolished.
12777 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12779 *Broadcom, tweaked and integrated by Geoff Thorpe*
12781 * Move common extension printing code to new function
12782 X509V3_print_extensions(). Reorganise OCSP print routines and
12783 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12787 * New function X509_signature_print() to remove duplication in some
12792 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12793 set (this was treated exactly the same as SET OF previously). This
12794 is used to reorder the STACK representing the structure to match the
12795 encoding. This will be used to get round a problem where a PKCS7
12796 structure which was signed could not be verified because the STACK
12797 order did not reflect the encoded order.
12801 * Reimplement the OCSP ASN1 module using the new code.
12805 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12806 for its ASN1 operations. The old style function pointers still exist
12807 for now but they will eventually go away.
12811 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12812 completely replaces the old ASN1 functionality with a table driven
12813 encoder and decoder which interprets an ASN1_ITEM structure describing
12814 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12815 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12816 has also been converted to the new form.
12820 * Change BN_mod_exp_recp so that negative moduli are tolerated
12821 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12822 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12823 for negative moduli.
12827 * Fix BN_uadd and BN_usub: Always return non-negative results instead
12828 of not touching the result's sign bit.
12832 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
12837 * Changed the LHASH code to use prototypes for callbacks, and created
12838 macros to declare and implement thin (optionally static) functions
12839 that provide type-safety and avoid function pointer casting for the
12840 type-specific callbacks.
12844 * Added Kerberos Cipher Suites to be used with TLS, as written in
12846 *Veers Staats <staatsvr@asc.hpc.mil>,
12847 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
12849 * Reformat the FAQ so the different questions and answers can be divided
12850 in sections depending on the subject.
12854 * Have the zlib compression code load ZLIB.DLL dynamically under
12859 * New function BN_mod_sqrt for computing square roots modulo a prime
12860 (using the probabilistic Tonelli-Shanks algorithm unless
12861 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
12862 be handled deterministically).
12864 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12866 * Make BN_mod_inverse faster by explicitly handling small quotients
12867 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
12868 512 bits], about 30% for larger ones [1024 or 2048 bits].)
12872 * New function BN_kronecker.
12876 * Fix BN_gcd so that it works on negative inputs; the result is
12877 positive unless both parameters are zero.
12878 Previously something reasonably close to an infinite loop was
12879 possible because numbers could be growing instead of shrinking
12880 in the implementation of Euclid's algorithm.
12884 * Fix BN_is_word() and BN_is_one() macros to take into account the
12885 sign of the number in question.
12887 Fix BN_is_word(a,w) to work correctly for w == 0.
12889 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
12890 because its test if the absolute value of 'a' equals 'w'.
12891 Note that BN_abs_is_word does *not* handle w == 0 reliably;
12892 it exists mostly for use in the implementations of BN_is_zero(),
12893 BN_is_one(), and BN_is_word().
12897 * New function BN_swap.
12901 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
12902 the exponentiation functions are more likely to produce reasonable
12903 results on negative inputs.
12907 * Change BN_mod_mul so that the result is always non-negative.
12908 Previously, it could be negative if one of the factors was negative;
12909 I don't think anyone really wanted that behaviour.
12913 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
12914 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
12915 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
12916 and add new functions:
12925 BN_mod_lshift1_quick
12927 BN_mod_lshift_quick
12929 These functions always generate non-negative results.
12931 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
12932 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
12934 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
12935 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
12936 be reduced modulo `m`.
12938 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
12941 The following entry accidentally appeared in the CHANGES file
12942 distributed with OpenSSL 0.9.7. The modifications described in
12943 it do *not* apply to OpenSSL 0.9.7.
12945 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
12946 was actually never needed) and in BN_mul(). The removal in BN_mul()
12947 required a small change in bn_mul_part_recursive() and the addition
12948 of the functions bn_cmp_part_words(), bn_sub_part_words() and
12949 bn_add_part_words(), which do the same thing as bn_cmp_words(),
12950 bn_sub_words() and bn_add_words() except they take arrays with
12956 * In 'openssl passwd', verify passwords read from the terminal
12957 unless the '-salt' option is used (which usually means that
12958 verification would just waste user's time since the resulting
12959 hash is going to be compared with some given password hash)
12960 or the new '-noverify' option is used.
12962 This is an incompatible change, but it does not affect
12963 non-interactive use of 'openssl passwd' (passwords on the command
12964 line, '-stdin' option, '-in ...' option) and thus should not
12965 cause any problems.
12969 * Remove all references to RSAref, since there's no more need for it.
12973 * Make DSO load along a path given through an environment variable
12974 (SHLIB_PATH) with shl_load().
12978 * Constify the ENGINE code as a result of BIGNUM constification.
12979 Also constify the RSA code and most things related to it. In a
12980 few places, most notable in the depth of the ASN.1 code, ugly
12981 casts back to non-const were required (to be solved at a later
12986 * Make it so the openssl application has all engines loaded by default.
12990 * Constify the BIGNUM routines a little more.
12994 * Add the following functions:
12996 ENGINE_load_cswift()
12998 ENGINE_load_atalla()
12999 ENGINE_load_nuron()
13000 ENGINE_load_builtin_engines()
13002 That way, an application can itself choose if external engines that
13003 are built-in in OpenSSL shall ever be used or not. The benefit is
13004 that applications won't have to be linked with libdl or other dso
13005 libraries unless it's really needed.
13007 Changed 'openssl engine' to load all engines on demand.
13008 Changed the engine header files to avoid the duplication of some
13009 declarations (they differed!).
13013 * 'openssl engine' can now list capabilities.
13017 * Better error reporting in 'openssl engine'.
13021 * Never call load_dh_param(NULL) in s_server.
13025 * Add engine application. It can currently list engines by name and
13026 identity, and test if they are actually available.
13030 * Improve RPM specification file by forcing symbolic linking and making
13031 sure the installed documentation is also owned by root.root.
13033 *Damien Miller <djm@mindrot.org>*
13035 * Give the OpenSSL applications more possibilities to make use of
13036 keys (public as well as private) handled by engines.
13040 * Add OCSP code that comes from CertCo.
13044 * Add VMS support for the Rijndael code.
13048 * Added untested support for Nuron crypto accelerator.
13052 * Add support for external cryptographic devices. This code was
13053 previously distributed separately as the "engine" branch.
13055 *Geoff Thorpe, Richard Levitte*
13057 * Rework the filename-translation in the DSO code. It is now possible to
13058 have far greater control over how a "name" is turned into a filename
13059 depending on the operating environment and any oddities about the
13060 different shared library filenames on each system.
13064 * Support threads on FreeBSD-elf in Configure.
13068 * Fix for SHA1 assembly problem with MASM: it produces
13069 warnings about corrupt line number information when assembling
13070 with debugging information. This is caused by the overlapping
13073 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13076 NCONF_get_number() has no error checking at all. As a replacement,
13077 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13078 promoted strongly. The old NCONF_get_number is kept around for
13079 binary backward compatibility.
13080 Make it possible for methods to load from something other than a BIO,
13081 by providing a function pointer that is given a name instead of a BIO.
13082 For example, this could be used to load configuration data from an
13087 * Fix for non blocking accept BIOs. Added new I/O special reason
13088 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13089 with non blocking I/O was not possible because no retry code was
13090 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13095 * Added the beginnings of Rijndael support.
13099 * Fix for bug in DirectoryString mask setting. Add support for
13100 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13101 to allow certificate printing to more controllable, additional
13102 'certopt' option to 'x509' to allow new printing options to be
13107 * Clean old EAY MD5 hack from e_os.h.
13111 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13113 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13114 by using the Codenomicon TLS Test Tool [CVE-2004-0079][]
13116 *Joe Orton, Steve Henson*
13118 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13120 * Fix additional bug revealed by the NISCC test suite:
13122 Stop bug triggering large recursion when presented with
13123 certain ASN.1 tags [CVE-2003-0851][]
13127 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13129 * Fix various bugs revealed by running the NISCC test suite:
13131 Stop out of bounds reads in the ASN1 code when presented with
13132 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13134 If verify callback ignores invalid public key errors don't try to check
13135 certificate signature with the NULL public key.
13139 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13140 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13145 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13146 extra data after the compression methods not only for TLS 1.0
13147 but also for SSL 3.0 (as required by the specification).
13149 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13151 * Change X509_certificate_type() to mark the key as exported/exportable
13152 when it's 512 *bits* long, not 512 bytes.
13156 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13158 * Countermeasure against the Klima-Pokorny-Rosa extension of
13159 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13160 a protocol version number mismatch like a decryption error
13161 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13165 * Turn on RSA blinding by default in the default implementation
13166 to avoid a timing attack. Applications that don't want it can call
13167 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13168 They would be ill-advised to do so in most cases.
13170 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13172 * Change RSA blinding code so that it works when the PRNG is not
13173 seeded (in this case, the secret RSA exponent is abused as
13174 an unpredictable seed -- if it is not unpredictable, there
13175 is no point in blinding anyway). Make RSA blinding thread-safe
13176 by remembering the creator's thread ID in rsa->blinding and
13177 having all other threads use local one-time blinding factors
13178 (this requires more computation than sharing rsa->blinding, but
13179 avoids excessive locking; and if an RSA object is not shared
13180 between threads, blinding will still be very fast).
13184 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13186 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13187 via timing by performing a MAC computation even if incorrect
13188 block cipher padding has been found. This is a countermeasure
13189 against active attacks where the attacker has to distinguish
13190 between bad padding and a MAC verification error. [CVE-2003-0078][]
13192 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13193 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13194 Martin Vuagnoux (EPFL, Ilion)*
13196 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13198 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13199 memory from its contents. This is done with a counter that will
13200 place alternating values in each byte. This can be used to solve
13201 two issues: 1) the removal of calls to memset() by highly optimizing
13202 compilers, and 2) cleansing with other values than 0, since those can
13203 be read through on certain media, for example a swap space on disk.
13207 * Bugfix: client side session caching did not work with external caching,
13208 because the session->cipher setting was not restored when reloading
13209 from the external cache. This problem was masked, when
13210 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13211 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13215 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13216 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13218 *Zeev Lieber <zeev-l@yahoo.com>*
13220 * Undo an undocumented change introduced in 0.9.6e which caused
13221 repeated calls to OpenSSL_add_all_ciphers() and
13222 OpenSSL_add_all_digests() to be ignored, even after calling
13227 * Change the default configuration reader to deal with last line not
13228 being properly terminated.
13232 * Change X509_NAME_cmp() so it applies the special rules on handling
13233 DN values that are of type PrintableString, as well as RDNs of type
13234 emailAddress where the value has the type ia5String.
13236 *stefank@valicert.com via Richard Levitte*
13238 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13239 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13240 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13241 the bitwise-OR of the two for use by the majority of applications
13242 wanting this behaviour, and update the docs. The documented
13243 behaviour and actual behaviour were inconsistent and had been
13244 changing anyway, so this is more a bug-fix than a behavioural
13247 *Geoff Thorpe, diagnosed by Nadav Har'El*
13249 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13250 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13254 * Fix initialization code race conditions in
13255 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13256 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13257 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13258 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13259 ssl2_get_cipher_by_char(),
13260 ssl3_get_cipher_by_char().
13262 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13264 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13265 the cached sessions are flushed, as the remove_cb() might use ex_data
13266 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13267 (see [openssl.org #212]).
13269 *Geoff Thorpe, Lutz Jaenicke*
13271 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13272 length, instead of the encoding length to d2i_ASN1_OBJECT.
13276 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13278 * [In 0.9.6g-engine release:]
13279 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13281 *Lynn Gazis <lgazis@rainbow.com>*
13283 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13285 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13286 and get fix the header length calculation.
13287 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13288 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13290 * Use proper error handling instead of 'assertions' in buffer
13291 overflow checks added in 0.9.6e. This prevents DoS (the
13292 assertions could call abort()).
13294 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13296 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13298 * Add various sanity checks to asn1_get_length() to reject
13299 the ASN1 length bytes if they exceed sizeof(long), will appear
13300 negative or the content length exceeds the length of the
13303 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13305 * Fix cipher selection routines: ciphers without encryption had no flags
13306 for the cipher strength set and where therefore not handled correctly
13307 by the selection routines (PR #130).
13311 * Fix EVP_dsa_sha macro.
13316 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13317 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13318 that was added in OpenSSL 0.9.6d.
13320 As the countermeasure turned out to be incompatible with some
13321 broken SSL implementations, the new option is part of SSL_OP_ALL.
13322 SSL_OP_ALL is usually employed when compatibility with weird SSL
13323 implementations is desired (e.g. '-bugs' option to 's_client' and
13324 's_server'), so the new option is automatically set in many
13329 * Changes in security patch:
13331 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13332 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13333 Air Force Materiel Command, USAF, under agreement number
13336 * Add various sanity checks to asn1_get_length() to reject
13337 the ASN1 length bytes if they exceed sizeof(long), will appear
13338 negative or the content length exceeds the length of the
13339 supplied buffer. [CVE-2002-0659][]
13341 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13343 * Assertions for various potential buffer overflows, not known to
13344 happen in practice.
13346 *Ben Laurie (CHATS)*
13348 * Various temporary buffers to hold ASCII versions of integers were
13349 too small for 64 bit platforms. [CVE-2002-0655][]
13350 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13352 * Remote buffer overflow in SSL3 protocol - an attacker could
13353 supply an oversized session ID to a client. [CVE-2002-0656][]
13355 *Ben Laurie (CHATS)*
13357 * Remote buffer overflow in SSL2 protocol - an attacker could
13358 supply an oversized client master key. [CVE-2002-0656][]
13360 *Ben Laurie (CHATS)*
13362 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13364 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13365 encoded as NULL) with id-dsa-with-sha1.
13367 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13369 * Check various `X509_...()` return values in apps/req.c.
13371 *Nils Larsch <nla@trustcenter.de>*
13373 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13374 an end-of-file condition would erroneously be flagged, when the CRLF
13375 was just at the end of a processed block. The bug was discovered when
13376 processing data through a buffering memory BIO handing the data to a
13377 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13378 <ptsekov@syntrex.com> and Nedelcho Stanev.
13382 * Implement a countermeasure against a vulnerability recently found
13383 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13384 before application data chunks to avoid the use of known IVs
13385 with data potentially chosen by the attacker.
13389 * Fix length checks in ssl3_get_client_hello().
13393 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13394 to prevent ssl3_read_internal() from incorrectly assuming that
13395 ssl3_read_bytes() found application data while handshake
13396 processing was enabled when in fact s->s3->in_read_app_data was
13397 merely automatically cleared during the initial handshake.
13399 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13401 * Fix object definitions for Private and Enterprise: they were not
13402 recognized in their shortname (=lowercase) representation. Extend
13403 obj_dat.pl to issue an error when using undefined keywords instead
13404 of silently ignoring the problem (Svenning Sorensen
13405 <sss@sss.dnsalias.net>).
13409 * Fix DH_generate_parameters() so that it works for 'non-standard'
13410 generators, i.e. generators other than 2 and 5. (Previously, the
13411 code did not properly initialise the 'add' and 'rem' values to
13412 BN_generate_prime().)
13414 In the new general case, we do not insist that 'generator' is
13415 actually a primitive root: This requirement is rather pointless;
13416 a generator of the order-q subgroup is just as good, if not
13421 * Map new X509 verification errors to alerts. Discovered and submitted by
13422 Tom Wu <tom@arcot.com>.
13426 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13427 returning non-zero before the data has been completely received
13428 when using non-blocking I/O.
13430 *Bodo Moeller; problem pointed out by John Hughes*
13432 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13434 *Ben Laurie, Lutz Jaenicke*
13436 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13437 Yoram Zahavi <YoramZ@gilian.com>).
13441 * Add information about CygWin 1.3 and on, and preserve proper
13442 configuration for the versions before that.
13444 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13446 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13447 check whether we deal with a copy of a session and do not delete from
13448 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13449 <izhar@checkpoint.com>.
13453 * Do not store session data into the internal session cache, if it
13454 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13455 flag is set). Proposed by Aslam <aslam@funk.com>.
13459 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13464 * [In 0.9.6d-engine release:]
13465 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13467 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13469 * Add the configuration target linux-s390x.
13471 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13473 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13474 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13475 variable as an indication that a ClientHello message has been
13476 received. As the flag value will be lost between multiple
13477 invocations of ssl3_accept when using non-blocking I/O, the
13478 function may not be aware that a handshake has actually taken
13479 place, thus preventing a new session from being added to the
13482 To avoid this problem, we now set s->new_session to 2 instead of
13483 using a local variable.
13485 *Lutz Jaenicke, Bodo Moeller*
13487 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13488 if the SSL_R_LENGTH_MISMATCH error is detected.
13490 *Geoff Thorpe, Bodo Moeller*
13492 * New 'shared_ldflag' column in Configure platform table.
13496 * Fix EVP_CIPHER_mode macro.
13498 *"Dan S. Camper" <dan@bti.net>*
13500 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13501 type, we must throw them away by setting rr->length to 0.
13503 *D P Chang <dpc@qualys.com>*
13505 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13507 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13508 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13509 worked incorrectly for those cases where range = `10..._2` and
13510 `3*range` is two bits longer than range.)
13514 * Only add signing time to PKCS7 structures if it is not already
13519 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13520 OBJ_ld_ce should be OBJ_id_ce.
13521 Also some ip-pda OIDs in crypto/objects/objects.txt were
13522 incorrect (cf. RFC 3039).
13524 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13526 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13527 returns early because it has nothing to do.
13529 *Andy Schneider <andy.schneider@bjss.co.uk>*
13531 * [In 0.9.6c-engine release:]
13532 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13534 *Andy Schneider <andy.schneider@bjss.co.uk>*
13536 * [In 0.9.6c-engine release:]
13537 Add support for Cryptographic Appliance's keyserver technology.
13538 (Use engine 'keyclient')
13540 *Cryptographic Appliances and Geoff Thorpe*
13542 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13543 is called via tools/c89.sh because arguments have to be
13544 rearranged (all '-L' options must appear before the first object
13547 *Richard Shapiro <rshapiro@abinitio.com>*
13549 * [In 0.9.6c-engine release:]
13550 Add support for Broadcom crypto accelerator cards, backported
13553 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13555 * [In 0.9.6c-engine release:]
13556 Add support for SureWare crypto accelerator cards from
13557 Baltimore Technologies. (Use engine 'sureware')
13559 *Baltimore Technologies and Mark Cox*
13561 * [In 0.9.6c-engine release:]
13562 Add support for crypto accelerator cards from Accelerated
13563 Encryption Processing, www.aep.ie. (Use engine 'aep')
13565 *AEP Inc. and Mark Cox*
13567 * Add a configuration entry for gcc on UnixWare.
13569 *Gary Benson <gbenson@redhat.com>*
13571 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13572 messages are stored in a single piece (fixed-length part and
13573 variable-length part combined) and fix various bugs found on the way.
13577 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13578 instead. BIO_gethostbyname() does not know what timeouts are
13579 appropriate, so entries would stay in cache even when they have
13581 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13583 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13584 faced with a pathologically small ClientHello fragment that does
13585 not contain client_version: Instead of aborting with an error,
13586 simply choose the highest available protocol version (i.e.,
13587 TLS 1.0 unless it is disabled). In practice, ClientHello
13588 messages are never sent like this, but this change gives us
13589 strictly correct behaviour at least for TLS.
13593 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13594 never resets s->method to s->ctx->method when called from within
13595 one of the SSL handshake functions.
13597 *Bodo Moeller; problem pointed out by Niko Baric*
13599 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13600 (sent using the client's version number) if client_version is
13601 smaller than the protocol version in use. Also change
13602 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13603 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13604 the client will at least see that alert.
13608 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13613 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13614 client receives HelloRequest while in a handshake.
13616 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13618 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13619 should end in 'break', not 'goto end' which circumvents various
13620 cleanups done in state SSL_ST_OK. But session related stuff
13621 must be disabled for SSL_ST_OK in the case that we just sent a
13624 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13625 before just sending a HelloRequest.
13627 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13629 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13630 reveal whether illegal block cipher padding was found or a MAC
13631 verification error occurred. (Neither SSLerr() codes nor alerts
13632 are directly visible to potential attackers, but the information
13633 may leak via logfiles.)
13635 Similar changes are not required for the SSL 2.0 implementation
13636 because the number of padding bytes is sent in clear for SSL 2.0,
13637 and the extra bytes are just ignored. However ssl/s2_pkt.c
13638 failed to verify that the purported number of padding bytes is in
13643 * Add OpenUNIX-8 support including shared libraries
13644 (Boyd Lynn Gerber <gerberb@zenez.com>).
13648 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13649 'wristwatch attack' using huge encoding parameters (cf.
13650 James H. Manger's CRYPTO 2001 paper). Note that the
13651 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13652 encoding parameters and hence was not vulnerable.
13656 * BN_sqr() bug fix.
13658 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13660 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13661 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13662 followed by modular reduction.
13664 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13666 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13667 equivalent based on BN_pseudo_rand() instead of BN_rand().
13671 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13672 This function was broken, as the check for a new client hello message
13673 to handle SGC did not allow these large messages.
13674 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13678 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13682 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13683 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13687 * Rework the configuration and shared library support for Tru64 Unix.
13688 The configuration part makes use of modern compiler features and
13689 still retains old compiler behavior for those that run older versions
13690 of the OS. The shared library support part includes a variant that
13691 uses the RPATH feature, and is available through the special
13692 configuration target "alpha-cc-rpath", which will never be selected
13695 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13697 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13698 with the same message size as in ssl3_get_certificate_request().
13699 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13700 messages might inadvertently be reject as too long.
13702 *Petr Lampa <lampa@fee.vutbr.cz>*
13704 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13708 * Modified SSL library such that the verify_callback that has been set
13709 specifically for an SSL object with SSL_set_verify() is actually being
13710 used. Before the change, a verify_callback set with this function was
13711 ignored and the verify_callback() set in the SSL_CTX at the time of
13712 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13713 to allow the necessary settings.
13717 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13718 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13719 done automatically (in contradiction to the requirements of the C
13720 standard). This made problems when used from OpenSSH.
13724 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13725 dh->length and always used
13727 BN_rand_range(priv_key, dh->p).
13729 BN_rand_range() is not necessary for Diffie-Hellman, and this
13730 specific range makes Diffie-Hellman unnecessarily inefficient if
13731 dh->length (recommended exponent length) is much smaller than the
13732 length of dh->p. We could use BN_rand_range() if the order of
13733 the subgroup was stored in the DH structure, but we only have
13738 BN_rand(priv_key, l, ...)
13740 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13747 RSA_eay_public_encrypt
13748 RSA_eay_private_decrypt
13749 RSA_eay_private_encrypt (signing)
13750 RSA_eay_public_decrypt (signature verification)
13752 (default implementations for RSA_public_encrypt,
13753 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13754 always reject numbers >= n.
13758 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13759 to synchronize access to 'locking_thread'. This is necessary on
13760 systems where access to 'locking_thread' (an 'unsigned long'
13761 variable) is not atomic.
13765 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13766 *before* setting the 'crypto_lock_rand' flag. The previous code had
13767 a race condition if 0 is a valid thread ID.
13769 *Travis Vitek <vitek@roguewave.com>*
13771 * Add support for shared libraries under Irix.
13773 *Albert Chin-A-Young <china@thewrittenword.com>*
13775 * Add configuration option to build on Linux on both big-endian and
13776 little-endian MIPS.
13778 *Ralf Baechle <ralf@uni-koblenz.de>*
13780 * Add the possibility to create shared libraries on HP-UX.
13784 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13786 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13787 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13788 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13789 PRNG state recovery was possible based on the output of
13790 one PRNG request appropriately sized to gain knowledge on
13791 'md' followed by enough consecutive 1-byte PRNG requests
13792 to traverse all of 'state'.
13794 1. When updating 'md_local' (the current thread's copy of 'md')
13795 during PRNG output generation, hash all of the previous
13796 'md_local' value, not just the half used for PRNG output.
13798 2. Make the number of bytes from 'state' included into the hash
13799 independent from the number of PRNG bytes requested.
13801 The first measure alone would be sufficient to avoid
13802 Markku-Juhani's attack. (Actually it had never occurred
13803 to me that the half of 'md_local' used for chaining was the
13804 half from which PRNG output bytes were taken -- I had always
13805 assumed that the secret half would be used.) The second
13806 measure makes sure that additional data from 'state' is never
13807 mixed into 'md_local' in small portions; this heuristically
13808 further strengthens the PRNG.
13812 * Fix crypto/bn/asm/mips3.s.
13816 * When only the key is given to "enc", the IV is undefined. Print out
13817 an error message in this case.
13821 * Handle special case when X509_NAME is empty in X509 printing routines.
13825 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
13826 positive and less than q.
13830 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
13831 used: it isn't thread safe and the add_lock_callback should handle
13834 *Paul Rose <Paul.Rose@bridge.com>*
13836 * Verify that incoming data obeys the block size in
13837 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
13843 *Ulf Möller, Bodo Möller*
13845 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
13846 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
13847 when fixing the server behaviour for backwards-compatible 'client
13848 hello' messages. (Note that the attack is impractical against
13849 SSL 3.0 and TLS 1.0 anyway because length and version checking
13850 means that the probability of guessing a valid ciphertext is
13851 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
13854 Before 0.9.5, the countermeasure (hide the error by generating a
13855 random 'decryption result') did not work properly because
13856 ERR_clear_error() was missing, meaning that SSL_get_error() would
13857 detect the supposedly ignored error.
13859 Both problems are now fixed.
13863 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
13864 (previously it was 1024).
13868 * Fix for compatibility mode trust settings: ignore trust settings
13869 unless some valid trust or reject settings are present.
13873 * Fix for blowfish EVP: its a variable length cipher.
13877 * Fix various bugs related to DSA S/MIME verification. Handle missing
13878 parameters in DSA public key structures and return an error in the
13879 DSA routines if parameters are absent.
13883 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
13884 in the current directory if neither $RANDFILE nor $HOME was set.
13885 RAND_file_name() in 0.9.6a returned NULL in this case. This has
13886 caused some confusion to Windows users who haven't defined $HOME.
13887 Thus RAND_file_name() is changed again: e_os.h can define a
13888 DEFAULT_HOME, which will be used if $HOME is not set.
13889 For Windows, we use "C:"; on other platforms, we still require
13890 environment variables.
13892 * Move 'if (!initialized) RAND_poll()' into regions protected by
13893 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
13894 having multiple threads call RAND_poll() concurrently.
13898 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
13899 combination of a flag and a thread ID variable.
13900 Otherwise while one thread is in ssleay_rand_bytes (which sets the
13901 flag), *other* threads can enter ssleay_add_bytes without obeying
13902 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
13903 that they do not hold after the first thread unsets add_do_not_lock).
13907 * Change bctest again: '-x' expressions are not available in all
13908 versions of 'test'.
13912 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
13914 * Fix a couple of memory leaks in PKCS7_dataDecode()
13916 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
13918 * Change Configure and Makefiles to provide EXE_EXT, which will contain
13919 the default extension for executables, if any. Also, make the perl
13920 scripts that use symlink() to test if it really exists and use "cp"
13921 if it doesn't. All this made OpenSSL compilable and installable in
13926 * Fix for asn1_GetSequence() for indefinite length constructed data.
13927 If SEQUENCE is length is indefinite just set c->slen to the total
13928 amount of data available.
13930 *Steve Henson, reported by shige@FreeBSD.org*
13932 *This change does not apply to 0.9.7.*
13934 * Change bctest to avoid here-documents inside command substitution
13935 (workaround for FreeBSD /bin/sh bug).
13936 For compatibility with Ultrix, avoid shell functions (introduced
13937 in the bctest version that searches along $PATH).
13941 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
13942 with des_encrypt() defined on some operating systems, like Solaris
13947 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
13948 On the Importance of Eliminating Errors in Cryptographic
13949 Computations, J. Cryptology 14 (2001) 2, 101-119,
13950 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
13954 * MIPS assembler BIGNUM division bug fix.
13958 * Disabled incorrect Alpha assembler code.
13962 * Fix PKCS#7 decode routines so they correctly update the length
13963 after reading an EOC for the EXPLICIT tag.
13967 *This change does not apply to 0.9.7.*
13969 * Fix bug in PKCS#12 key generation routines. This was triggered
13970 if a 3DES key was generated with a 0 initial byte. Include
13971 PKCS12_BROKEN_KEYGEN compilation option to retain the old
13972 (but broken) behaviour.
13976 * Enhance bctest to search for a working bc along $PATH and print
13979 *Tim Rice <tim@multitalents.net> via Richard Levitte*
13981 * Fix memory leaks in err.c: free err_data string if necessary;
13982 don't write to the wrong index in ERR_set_error_data.
13986 * Implement ssl23_peek (analogous to ssl23_read), which previously
13991 * Replace rdtsc with `_emit` statements for VC++ version 5.
13993 *Jeremy Cooper <jeremy@baymoo.org>*
13995 * Make it possible to reuse SSLv2 sessions.
13999 * In copy_email() check for >= 0 as a return value for
14000 X509_NAME_get_index_by_NID() since 0 is a valid index.
14002 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14004 * Avoid coredump with unsupported or invalid public keys by checking if
14005 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14006 PKCS7_verify() fails with non detached data.
14010 * Don't use getenv in library functions when run as setuid/setgid.
14011 New function OPENSSL_issetugid().
14015 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14016 due to incorrect handling of multi-threading:
14018 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14020 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14022 3. Count how many times MemCheck_off() has been called so that
14023 nested use can be treated correctly. This also avoids
14024 inband-signalling in the previous code (which relied on the
14025 assumption that thread ID 0 is impossible).
14029 * Add "-rand" option also to s_client and s_server.
14033 * Fix CPU detection on Irix 6.x.
14034 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14035 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14037 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14042 *This change does not apply to 0.9.7.*
14044 * Use the cached encoding of an X509_NAME structure rather than
14045 copying it. This is apparently the reason for the libsafe "errors"
14046 but the code is actually correct.
14050 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14051 Bleichenbacher's DSA attack.
14052 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14053 to be set and top=0 forces the highest bit to be set; top=-1 is new
14054 and leaves the highest bit random.
14056 *Ulf Moeller, Bodo Moeller*
14058 * In the `NCONF_...`-based implementations for `CONF_...` queries
14059 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14060 a temporary CONF structure with the data component set to NULL
14061 (which gives segmentation faults in lh_retrieve).
14062 Instead, use NULL for the CONF pointer in CONF_get_string and
14063 CONF_get_number (which may use environment variables) and directly
14064 return NULL from CONF_get_section.
14068 * Fix potential buffer overrun for EBCDIC.
14072 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14073 keyUsage if basicConstraints absent for a CA.
14077 * Make SMIME_write_PKCS7() write mail header values with a format that
14078 is more generally accepted (no spaces before the semicolon), since
14079 some programs can't parse those values properly otherwise. Also make
14080 sure BIO's that break lines after each write do not create invalid
14085 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14086 macros previously used would not encode an empty SEQUENCE OF
14087 and break the signature.
14091 *This change does not apply to 0.9.7.*
14093 * Zero the premaster secret after deriving the master secret in
14098 * Add some EVP_add_digest_alias registrations (as found in
14099 OpenSSL_add_all_digests()) to SSL_library_init()
14100 aka OpenSSL_add_ssl_algorithms(). This provides improved
14101 compatibility with peers using X.509 certificates
14102 with unconventional AlgorithmIdentifier OIDs.
14106 * Fix for Irix with NO_ASM.
14108 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14110 * ./config script fixes.
14112 *Ulf Moeller, Richard Levitte*
14114 * Fix 'openssl passwd -1'.
14118 * Change PKCS12_key_gen_asc() so it can cope with non null
14119 terminated strings whose length is passed in the passlen
14120 parameter, for example from PEM callbacks. This was done
14121 by adding an extra length parameter to asc2uni().
14123 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14125 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14126 call failed, free the DSA structure.
14130 * Fix to uni2asc() to cope with zero length Unicode strings.
14131 These are present in some PKCS#12 files.
14135 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14136 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14137 when writing a 32767 byte record.
14139 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14141 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14142 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14144 (RSA objects have a reference count access to which is protected
14145 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14146 so they are meant to be shared between threads.)
14147 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14148 "Reddie, Steven" <Steven.Reddie@ca.com>*
14150 * Fix a deadlock in CRYPTO_mem_leaks().
14154 * Use better test patterns in bntest.
14158 * rand_win.c fix for Borland C.
14162 * BN_rshift bugfix for n == 0.
14166 * Add a 'bctest' script that checks for some known 'bc' bugs
14167 so that 'make test' does not abort just because 'bc' is broken.
14171 * Store verify_result within SSL_SESSION also for client side to
14172 avoid potential security hole. (Re-used sessions on the client side
14173 always resulted in verify_result==X509_V_OK, not using the original
14174 result of the server certificate verification.)
14178 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14179 SSL3_RT_APPLICATION_DATA, return 0.
14180 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14185 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14186 releases, have been re-implemented by renaming the previous
14187 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14188 and ssl3_read_internal, respectively, and adding 'peek' parameters
14189 to them. The new ssl[23]_{read,peek} functions are calls to
14190 ssl[23]_read_internal with the 'peek' flag set appropriately.
14191 A 'peek' parameter has also been added to ssl3_read_bytes, which
14192 does the actual work for ssl3_read_internal.
14196 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14197 the method-specific "init()" handler. Also clean up ex_data after
14198 calling the method-specific "finish()" handler. Previously, this was
14199 happening the other way round.
14203 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14204 The previous value, 12, was not always sufficient for BN_mod_exp().
14208 * Make sure that shared libraries get the internal name engine with
14209 the full version number and not just 0. This should mark the
14210 shared libraries as not backward compatible. Of course, this should
14211 be changed again when we can guarantee backward binary compatibility.
14215 * Fix typo in get_cert_by_subject() in by_dir.c
14217 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14219 * Rework the system to generate shared libraries:
14221 - Make note of the expected extension for the shared libraries and
14222 if there is a need for symbolic links from for example libcrypto.so.0
14223 to libcrypto.so.0.9.7. There is extended info in Configure for
14226 - Make as few rebuilds of the shared libraries as possible.
14228 - Still avoid linking the OpenSSL programs with the shared libraries.
14230 - When installing, install the shared libraries separately from the
14235 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14237 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14238 and not in SSL_clear because the latter is also used by the
14239 accept/connect functions; previously, the settings made by
14240 SSL_set_read_ahead would be lost during the handshake.
14242 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14244 * Correct util/mkdef.pl to be selective about disabled algorithms.
14245 Previously, it would create entries for disabled algorithms no
14250 * Added several new manual pages for SSL_* function.
14254 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14256 * In ssl23_get_client_hello, generate an error message when faced
14257 with an initial SSL 3.0/TLS record that is too small to contain the
14258 first two bytes of the ClientHello message, i.e. client_version.
14259 (Note that this is a pathologic case that probably has never happened
14260 in real life.) The previous approach was to use the version number
14261 from the record header as a substitute; but our protocol choice
14262 should not depend on that one because it is not authenticated
14263 by the Finished messages.
14267 * More robust randomness gathering functions for Windows.
14269 *Jeffrey Altman <jaltman@columbia.edu>*
14271 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14272 not set then we don't setup the error code for issuer check errors
14273 to avoid possibly overwriting other errors which the callback does
14274 handle. If an application does set the flag then we assume it knows
14275 what it is doing and can handle the new informational codes
14280 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14281 a general "ANY" type, as such it should be able to decode anything
14282 including tagged types. However it didn't check the class so it would
14283 wrongly interpret tagged types in the same way as their universal
14284 counterpart and unknown types were just rejected. Changed so that the
14285 tagged and unknown types are handled in the same way as a SEQUENCE:
14286 that is the encoding is stored intact. There is also a new type
14287 "V_ASN1_OTHER" which is used when the class is not universal, in this
14288 case we have no idea what the actual type is so we just lump them all
14293 * On VMS, stdout may very well lead to a file that is written to
14294 in a record-oriented fashion. That means that every write() will
14295 write a separate record, which will be read separately by the
14296 programs trying to read from it. This can be very confusing.
14298 The solution is to put a BIO filter in the way that will buffer
14299 text until a linefeed is reached, and then write everything a
14300 line at a time, so every record written will be an actual line,
14301 not chunks of lines and not (usually doesn't happen, but I've
14302 seen it once) several lines in one record. BIO_f_linebuffer() is
14305 Currently, it's a VMS-only method, because that's where it has
14306 been tested well enough.
14310 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14311 it can return incorrect results.
14312 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14313 but it was in 0.9.6-beta[12].)
14317 * Disable the check for content being present when verifying detached
14318 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14319 include zero length content when signing messages.
14323 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14324 BIO_ctrl (for BIO pairs).
14328 * Add DSO method for VMS.
14332 * Bug fix: Montgomery multiplication could produce results with the
14337 * Add RPM specification openssl.spec and modify it to build three
14338 packages. The default package contains applications, application
14339 documentation and run-time libraries. The devel package contains
14340 include files, static libraries and function documentation. The
14341 doc package contains the contents of the doc directory. The original
14342 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14346 * Add a large number of documentation files for many SSL routines.
14348 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14350 * Add a configuration entry for Sony News 4.
14352 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14354 * Don't set the two most significant bits to one when generating a
14355 random number < q in the DSA library.
14359 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14360 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14361 the underlying transport is blocking) if a handshake took place.
14362 (The default behaviour is needed by applications such as s_client
14363 and s_server that use select() to determine when to use SSL_read;
14364 but for applications that know in advance when to expect data, it
14365 just makes things more complicated.)
14369 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14374 * Add a few more EBCDIC conditionals that make `req` and `x509`
14375 work better on such systems.
14377 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14379 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14380 Update PKCS12_parse() so it copies the friendlyName and the
14381 keyid to the certificates aux info.
14385 * Fix bug in PKCS7_verify() which caused an infinite loop
14386 if there was more than one signature.
14388 *Sven Uszpelkat <su@celocom.de>*
14390 * Major change in util/mkdef.pl to include extra information
14391 about each symbol, as well as presenting variables as well
14392 as functions. This change means that there's n more need
14393 to rebuild the .num files when some algorithms are excluded.
14397 * Allow the verify time to be set by an application,
14398 rather than always using the current time.
14402 * Phase 2 verify code reorganisation. The certificate
14403 verify code now looks up an issuer certificate by a
14404 number of criteria: subject name, authority key id
14405 and key usage. It also verifies self signed certificates
14406 by the same criteria. The main comparison function is
14407 X509_check_issued() which performs these checks.
14409 Lot of changes were necessary in order to support this
14410 without completely rewriting the lookup code.
14412 Authority and subject key identifier are now cached.
14414 The LHASH 'certs' is X509_STORE has now been replaced
14415 by a STACK_OF(X509_OBJECT). This is mainly because an
14416 LHASH can't store or retrieve multiple objects with
14417 the same hash value.
14419 As a result various functions (which were all internal
14420 use only) have changed to handle the new X509_STORE
14421 structure. This will break anything that messed round
14422 with X509_STORE internally.
14424 The functions X509_STORE_add_cert() now checks for an
14425 exact match, rather than just subject name.
14427 The X509_STORE API doesn't directly support the retrieval
14428 of multiple certificates matching a given criteria, however
14429 this can be worked round by performing a lookup first
14430 (which will fill the cache with candidate certificates)
14431 and then examining the cache for matches. This is probably
14432 the best we can do without throwing out X509_LOOKUP
14433 entirely (maybe later...).
14435 The X509_VERIFY_CTX structure has been enhanced considerably.
14437 All certificate lookup operations now go via a get_issuer()
14438 callback. Although this currently uses an X509_STORE it
14439 can be replaced by custom lookups. This is a simple way
14440 to bypass the X509_STORE hackery necessary to make this
14441 work and makes it possible to use more efficient techniques
14442 in future. A very simple version which uses a simple
14443 STACK for its trusted certificate store is also provided
14444 using X509_STORE_CTX_trusted_stack().
14446 The verify_cb() and verify() callbacks now have equivalents
14447 in the X509_STORE_CTX structure.
14449 X509_STORE_CTX also has a 'flags' field which can be used
14450 to customise the verify behaviour.
14454 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14455 excludes S/MIME capabilities.
14459 * When a certificate request is read in keep a copy of the
14460 original encoding of the signed data and use it when outputting
14461 again. Signatures then use the original encoding rather than
14462 a decoded, encoded version which may cause problems if the
14463 request is improperly encoded.
14467 * For consistency with other BIO_puts implementations, call
14468 buffer_write(b, ...) directly in buffer_puts instead of calling
14471 In BIO_puts, increment b->num_write as in BIO_write.
14473 *Peter.Sylvester@EdelWeb.fr*
14475 * Fix BN_mul_word for the case where the word is 0. (We have to use
14476 BN_zero, we may not return a BIGNUM with an array consisting of
14477 words set to zero.)
14481 * Avoid calling abort() from within the library when problems are
14482 detected, except if preprocessor symbols have been defined
14483 (such as REF_CHECK, BN_DEBUG etc.).
14487 * New openssl application 'rsautl'. This utility can be
14488 used for low level RSA operations. DER public key
14489 BIO/fp routines also added.
14493 * New Configure entry and patches for compiling on QNX 4.
14495 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14497 * A demo state-machine implementation was sponsored by
14498 Nuron (<http://www.nuron.com/>) and is now available in
14499 demos/state_machine.
14503 * New options added to the 'dgst' utility for signature
14504 generation and verification.
14508 * Unrecognized PKCS#7 content types are now handled via a
14509 catch all ASN1_TYPE structure. This allows unsupported
14510 types to be stored as a "blob" and an application can
14511 encode and decode it manually.
14515 * Fix various signed/unsigned issues to make a_strex.c
14516 compile under VC++.
14518 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14520 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14521 length if passed a buffer. ASN1_INTEGER_to_BN failed
14522 if passed a NULL BN and its argument was negative.
14524 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14526 * Modification to PKCS#7 encoding routines to output definite
14527 length encoding. Since currently the whole structures are in
14528 memory there's not real point in using indefinite length
14529 constructed encoding. However if OpenSSL is compiled with
14530 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14534 * Added BIO_vprintf() and BIO_vsnprintf().
14538 * Added more prefixes to parse for in the strings written
14539 through a logging bio, to cover all the levels that are available
14540 through syslog. The prefixes are now:
14542 PANIC, EMERG, EMR => LOG_EMERG
14543 ALERT, ALR => LOG_ALERT
14544 CRIT, CRI => LOG_CRIT
14545 ERROR, ERR => LOG_ERR
14546 WARNING, WARN, WAR => LOG_WARNING
14547 NOTICE, NOTE, NOT => LOG_NOTICE
14548 INFO, INF => LOG_INFO
14549 DEBUG, DBG => LOG_DEBUG
14551 and as before, if none of those prefixes are present at the
14552 beginning of the string, LOG_ERR is chosen.
14554 On Win32, the `LOG_*` levels are mapped according to this:
14556 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14557 LOG_WARNING => EVENTLOG_WARNING_TYPE
14558 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14562 * Made it possible to reconfigure with just the configuration
14563 argument "reconf" or "reconfigure". The command line arguments
14564 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14565 and are retrieved from there when reconfiguring.
14571 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14573 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14577 * The obj_dat.pl script was messing up the sorting of object
14578 names. The reason was that it compared the quoted version
14579 of strings as a result "OCSP" > "OCSP Signing" because
14580 " > SPACE. Changed script to store unquoted versions of
14581 names and add quotes on output. It was also omitting some
14582 names from the lookup table if they were given a default
14583 value (that is if SN is missing it is given the same
14584 value as LN and vice versa), these are now added on the
14585 grounds that if an object has a name we should be able to
14586 look it up. Finally added warning output when duplicate
14587 short or long names are found.
14591 * Changes needed for Tandem NSK.
14593 *Scott Uroff <scott@xypro.com>*
14595 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14596 RSA_padding_check_SSLv23(), special padding was never detected
14597 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14598 version rollback attacks was not effective.
14600 In s23_clnt.c, don't use special rollback-attack detection padding
14601 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14602 client; similarly, in s23_srvr.c, don't do the rollback check if
14603 SSL 2.0 is the only protocol enabled in the server.
14607 * Make it possible to get hexdumps of unprintable data with 'openssl
14608 asn1parse'. By implication, the functions ASN1_parse_dump() and
14609 BIO_dump_indent() are added.
14613 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14614 these print out strings and name structures based on various
14615 flags including RFC2253 support and proper handling of
14616 multibyte characters. Added options to the 'x509' utility
14617 to allow the various flags to be set.
14621 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14622 Also change the functions X509_cmp_current_time() and
14623 X509_gmtime_adj() work with an ASN1_TIME structure,
14624 this will enable certificates using GeneralizedTime in validity
14625 dates to be checked.
14629 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14630 negative public key encodings) on by default,
14631 NO_NEG_PUBKEY_BUG can be set to disable it.
14635 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14636 content octets. An i2c_ASN1_OBJECT is unnecessary because
14637 the encoding can be trivially obtained from the structure.
14641 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14642 not read locks (`CRYPTO_r_[un]lock`).
14646 * A first attempt at creating official support for shared
14647 libraries through configuration. I've kept it so the
14648 default is static libraries only, and the OpenSSL programs
14649 are always statically linked for now, but there are
14650 preparations for dynamic linking in place.
14651 This has been tested on Linux and Tru64.
14655 * Randomness polling function for Win9x, as described in:
14656 Peter Gutmann, Software Generation of Practically Strong
14661 * Fix so PRNG is seeded in req if using an already existing
14666 * New options to smime application. -inform and -outform
14667 allow alternative formats for the S/MIME message including
14668 PEM and DER. The -content option allows the content to be
14669 specified separately. This should allow things like Netscape
14670 form signing output easier to verify.
14674 * Fix the ASN1 encoding of tags using the 'long form'.
14678 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14679 STRING types. These convert content octets to and from the
14680 underlying type. The actual tag and length octets are
14681 already assumed to have been read in and checked. These
14682 are needed because all other string types have virtually
14683 identical handling apart from the tag. By having versions
14684 of the ASN1 functions that just operate on content octets
14685 IMPLICIT tagging can be handled properly. It also allows
14686 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14687 and ASN1_INTEGER are identical apart from the tag.
14691 * Change the handling of OID objects as follows:
14693 - New object identifiers are inserted in objects.txt, following
14694 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
14695 - objects.pl is used to process obj_mac.num and create a new
14697 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14700 This is currently kind of a hack, and the perl code in objects.pl
14701 isn't very elegant, but it works as I intended. The simplest way
14702 to check that it worked correctly is to look in obj_dat.h and
14703 check the array nid_objs and make sure the objects haven't moved
14704 around (this is important!). Additions are OK, as well as
14705 consistent name changes.
14709 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14713 * Addition of the command line parameter '-rand file' to 'openssl req'.
14714 The given file adds to whatever has already been seeded into the
14715 random pool through the RANDFILE configuration file option or
14716 environment variable, or the default random state file.
14720 * mkstack.pl now sorts each macro group into lexical order.
14721 Previously the output order depended on the order the files
14722 appeared in the directory, resulting in needless rewriting
14727 * Patches to make OpenSSL compile under Win32 again. Mostly
14728 work arounds for the VC++ problem that it treats func() as
14729 func(void). Also stripped out the parts of mkdef.pl that
14730 added extra typesafe functions: these no longer exist.
14734 * Reorganisation of the stack code. The macros are now all
14735 collected in safestack.h . Each macro is defined in terms of
14736 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14737 DEBUG_SAFESTACK is now handled in terms of function casts,
14738 this has the advantage of retaining type safety without the
14739 use of additional functions. If DEBUG_SAFESTACK is not defined
14740 then the non typesafe macros are used instead. Also modified the
14741 mkstack.pl script to handle the new form. Needs testing to see
14742 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14743 the default if no major problems. Similar behaviour for ASN1_SET_OF
14744 and PKCS12_STACK_OF.
14748 * When some versions of IIS use the 'NET' form of private key the
14749 key derivation algorithm is different. Normally MD5(password) is
14750 used as a 128 bit RC4 key. In the modified case
14751 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14752 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14753 as the old Netscape_RSA functions except they have an additional
14754 'sgckey' parameter which uses the modified algorithm. Also added
14755 an -sgckey command line option to the rsa utility. Thanks to
14756 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14757 algorithm to openssl-dev.
14761 * The evp_local.h macros were using 'c.##kname' which resulted in
14762 invalid expansion on some systems (SCO 5.0.5 for example).
14763 Corrected to 'c.kname'.
14765 *Phillip Porch <root@theporch.com>*
14767 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14768 a STACK of email addresses from a certificate or request, these look
14769 in the subject name and the subject alternative name extensions and
14770 omit any duplicate addresses.
14774 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14775 This makes DSA verification about 2 % faster.
14779 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14780 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14781 plus overhead for 1024 bit moduli).
14782 This makes exponentiations about 0.5 % faster for 1024 bit
14783 exponents (as measured by "openssl speed rsa2048").
14787 * Rename memory handling macros to avoid conflicts with other
14789 Malloc => OPENSSL_malloc
14790 Malloc_locked => OPENSSL_malloc_locked
14791 Realloc => OPENSSL_realloc
14792 Free => OPENSSL_free
14796 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14797 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14801 * CygWin32 support.
14803 *John Jarvie <jjarvie@newsguy.com>*
14805 * The type-safe stack code has been rejigged. It is now only compiled
14806 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14807 by default all type-specific stack functions are "#define"d back to
14808 standard stack functions. This results in more streamlined output
14809 but retains the type-safety checking possibilities of the original
14814 * The STACK code has been cleaned up, and certain type declarations
14815 that didn't make a lot of sense have been brought in line. This has
14816 also involved a cleanup of sorts in safestack.h to more correctly
14817 map type-safe stack functions onto their plain stack counterparts.
14818 This work has also resulted in a variety of "const"ifications of
14819 lots of the code, especially `_cmp` operations which should normally
14820 be prototyped with "const" parameters anyway.
14824 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14825 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
14826 (The PRNG state consists of two parts, the large pool 'state' and 'md',
14827 where all of 'md' is used each time the PRNG is used, but 'state'
14828 is used only indexed by a cyclic counter. As entropy may not be
14829 well distributed from the beginning, 'md' is important as a
14830 chaining variable. However, the output function chains only half
14831 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
14832 all of 'md', and seeding with STATE_SIZE dummy bytes will result
14833 in all of 'state' being rewritten, with the new values depending
14834 on virtually all of 'md'. This overcomes the 80 bit limitation.)
14838 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
14839 the handshake is continued after ssl_verify_cert_chain();
14840 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
14841 can lead to 'unexplainable' connection aborts later.
14843 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
14845 * Major EVP API cipher revision.
14846 Add hooks for extra EVP features. This allows various cipher
14847 parameters to be set in the EVP interface. Support added for variable
14848 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
14849 setting of RC2 and RC5 parameters.
14851 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
14854 Remove lots of duplicated code from the EVP library. For example *every*
14855 cipher init() function handles the 'iv' in the same way according to the
14856 cipher mode. They also all do nothing if the 'key' parameter is NULL and
14857 for CFB and OFB modes they zero ctx->num.
14859 New functionality allows removal of S/MIME code RC2 hack.
14861 Most of the routines have the same form and so can be declared in terms
14864 By shifting this to the top level EVP_CipherInit() it can be removed from
14865 all individual ciphers. If the cipher wants to handle IVs or keys
14866 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
14869 Change lots of functions like EVP_EncryptUpdate() to now return a
14870 value: although software versions of the algorithms cannot fail
14871 any installed hardware versions can.
14875 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
14876 this option is set, tolerate broken clients that send the negotiated
14877 protocol version number instead of the requested protocol version
14882 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
14883 i.e. non-zero for export ciphersuites, zero otherwise.
14884 Previous versions had this flag inverted, inconsistent with
14885 rsa_tmp_cb (..._TMP_RSA_CB).
14887 *Bodo Moeller; problem reported by Amit Chopra*
14889 * Add missing DSA library text string. Work around for some IIS
14890 key files with invalid SEQUENCE encoding.
14894 * Add a document (doc/standards.txt) that list all kinds of standards
14895 and so on that are implemented in OpenSSL.
14899 * Enhance c_rehash script. Old version would mishandle certificates
14900 with the same subject name hash and wouldn't handle CRLs at all.
14901 Added -fingerprint option to crl utility, to support new c_rehash
14906 * Eliminate non-ANSI declarations in crypto.h and stack.h.
14910 * Fix for SSL server purpose checking. Server checking was
14911 rejecting certificates which had extended key usage present
14912 but no ssl client purpose.
14914 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
14916 * Make PKCS#12 code work with no password. The PKCS#12 spec
14917 is a little unclear about how a blank password is handled.
14918 Since the password in encoded as a BMPString with terminating
14919 double NULL a zero length password would end up as just the
14920 double NULL. However no password at all is different and is
14921 handled differently in the PKCS#12 key generation code. NS
14922 treats a blank password as zero length. MSIE treats it as no
14923 password on export: but it will try both on import. We now do
14924 the same: PKCS12_parse() tries zero length and no password if
14925 the password is set to "" or NULL (NULL is now a valid password:
14926 it wasn't before) as does the pkcs12 application.
14930 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
14931 perror when PEM_read_bio_X509_REQ fails, the error message must
14932 be obtained from the error queue.
14936 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
14937 it in ERR_remove_state if appropriate, and change ERR_get_state
14938 accordingly to avoid race conditions (this is necessary because
14939 thread_hash is no longer constant once set).
14943 * Bugfix for linux-elf makefile.one.
14947 * RSA_get_default_method() will now cause a default
14948 RSA_METHOD to be chosen if one doesn't exist already.
14949 Previously this was only set during a call to RSA_new()
14950 or RSA_new_method(NULL) meaning it was possible for
14951 RSA_get_default_method() to return NULL.
14955 * Added native name translation to the existing DSO code
14956 that will convert (if the flag to do so is set) filenames
14957 that are sufficiently small and have no path information
14958 into a canonical native form. Eg. "blah" converted to
14959 "libblah.so" or "blah.dll" etc.
14963 * New function ERR_error_string_n(e, buf, len) which is like
14964 ERR_error_string(e, buf), but writes at most 'len' bytes
14965 including the 0 terminator. For ERR_error_string_n, 'buf'
14968 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
14970 * CONF library reworked to become more general. A new CONF
14971 configuration file reader "class" is implemented as well as a
14972 new functions (`NCONF_*`, for "New CONF") to handle it. The now
14973 old `CONF_*` functions are still there, but are reimplemented to
14974 work in terms of the new functions. Also, a set of functions
14975 to handle the internal storage of the configuration data is
14976 provided to make it easier to write new configuration file
14977 reader "classes" (I can definitely see something reading a
14978 configuration file in XML format, for example), called `_CONF_*`,
14979 or "the configuration storage API"...
14981 The new configuration file reading functions are:
14983 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
14984 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
14986 NCONF_default, NCONF_WIN32
14988 NCONF_dump_fp, NCONF_dump_bio
14990 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
14991 NCONF_new creates a new CONF object. This works in the same way
14992 as other interfaces in OpenSSL, like the BIO interface.
14993 `NCONF_dump_*` dump the internal storage of the configuration file,
14994 which is useful for debugging. All other functions take the same
14995 arguments as the old `CONF_*` functions with the exception of the
14996 first that must be a `CONF *` instead of a `LHASH *`.
14998 To make it easier to use the new classes with the old `CONF_*` functions,
14999 the function CONF_set_default_method is provided.
15003 * Add '-tls1' option to 'openssl ciphers', which was already
15004 mentioned in the documentation but had not been implemented.
15005 (This option is not yet really useful because even the additional
15006 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15010 * Initial DSO code added into libcrypto for letting OpenSSL (and
15011 OpenSSL-based applications) load shared libraries and bind to
15012 them in a portable way.
15014 *Geoff Thorpe, with contributions from Richard Levitte*
15016 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15018 * Make sure _lrotl and _lrotr are only used with MSVC.
15020 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15021 (the default implementation of RAND_status).
15023 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15024 to '-clrext' (= clear extensions), as intended and documented.
15025 *Bodo Moeller; inconsistency pointed out by Michael Attili
15026 <attili@amaxo.com>*
15028 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15029 was larger than the MD block size.
15031 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15033 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15034 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15035 using the passed key: if the passed key was a private key the result
15036 of X509_print(), for example, would be to print out all the private key
15041 * des_quad_cksum() byte order bug fix.
15042 *Ulf Möller, using the problem description in krb4-0.9.7, where
15043 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15045 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15048 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15050 * For easily testing in shell scripts whether some command
15051 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15052 returns with exit code 0 iff no command of the given name is available.
15053 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15054 the output goes to stdout and nothing is printed to stderr.
15055 Additional arguments are always ignored.
15057 Since for each cipher there is a command of the same name,
15058 the 'no-cipher' compilation switches can be tested this way.
15060 ('openssl no-XXX' is not able to detect pseudo-commands such
15061 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15065 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15069 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15070 is set; it will be thrown away anyway because each handshake creates
15072 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15073 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15074 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15075 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15079 * New s_client option -ign_eof: EOF at stdin is ignored, and
15080 'Q' and 'R' lose their special meanings (quit/renegotiate).
15081 This is part of what -quiet does; unlike -quiet, -ign_eof
15082 does not suppress any output.
15086 * Add compatibility options to the purpose and trust code. The
15087 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15088 accepts a certificate or CA, this was the previous behaviour,
15089 with all the associated security issues.
15091 X509_TRUST_COMPAT is the old trust behaviour: only and
15092 automatically trust self signed roots in certificate store. A
15093 new trust setting X509_TRUST_DEFAULT is used to specify that
15094 a purpose has no associated trust setting and it should instead
15095 use the value in the default purpose.
15099 * Fix the PKCS#8 DSA private key code so it decodes keys again
15100 and fix a memory leak.
15104 * In util/mkerr.pl (which implements 'make errors'), preserve
15105 reason strings from the previous version of the .c file, as
15106 the default to have only downcase letters (and digits) in
15107 automatically generated reasons codes is not always appropriate.
15111 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15112 using strerror. Previously, ERR_reason_error_string() returned
15113 library names as reason strings for SYSerr; but SYSerr is a special
15114 case where small numbers are errno values, not library numbers.
15118 * Add '-dsaparam' option to 'openssl dhparam' application. This
15119 converts DSA parameters into DH parameters. (When creating parameters,
15120 DSA_generate_parameters is used.)
15124 * Include 'length' (recommended exponent length) in C code generated
15125 by 'openssl dhparam -C'.
15129 * The second argument to set_label in perlasm was already being used
15130 so couldn't be used as a "file scope" flag. Moved to third argument
15135 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15136 instead of RAND_bytes for encryption IVs and salts.
15140 * Include RAND_status() into RAND_METHOD instead of implementing
15141 it only for md_rand.c Otherwise replacing the PRNG by calling
15142 RAND_set_rand_method would be impossible.
15146 * Don't let DSA_generate_key() enter an infinite loop if the random
15147 number generation fails.
15151 * New 'rand' application for creating pseudo-random output.
15155 * Added configuration support for Linux/IA64
15157 *Rolf Haberrecker <rolf@suse.de>*
15159 * Assembler module support for Mingw32.
15163 * Shared library support for HPUX (in shlib/).
15165 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15167 * Shared library support for Solaris gcc.
15169 *Lutz Behnke <behnke@trustcenter.de>*
15171 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15173 * PKCS7_encrypt() was adding text MIME headers twice because they
15174 were added manually and by SMIME_crlf_copy().
15178 * In bntest.c don't call BN_rand with zero bits argument.
15180 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15182 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15183 case was implemented. This caused BN_div_recp() to fail occasionally.
15187 * Add an optional second argument to the set_label() in the perl
15188 assembly language builder. If this argument exists and is set
15189 to 1 it signals that the assembler should use a symbol whose
15190 scope is the entire file, not just the current function. This
15191 is needed with MASM which uses the format label:: for this scope.
15193 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15195 * Change the ASN1 types so they are typedefs by default. Before
15196 almost all types were #define'd to ASN1_STRING which was causing
15197 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15202 * Change names of new functions to the new get1/get0 naming
15203 convention: After 'get1', the caller owns a reference count
15204 and has to call `..._free`; 'get0' returns a pointer to some
15205 data structure without incrementing reference counters.
15206 (Some of the existing 'get' functions increment a reference
15207 counter, some don't.)
15208 Similarly, 'set1' and 'add1' functions increase reference
15209 counters or duplicate objects.
15213 * Allow for the possibility of temp RSA key generation failure:
15214 the code used to assume it always worked and crashed on failure.
15218 * Fix potential buffer overrun problem in BIO_printf().
15219 *Ulf Möller, using public domain code by Patrick Powell; problem
15220 pointed out by David Sacerdote <das33@cornell.edu>*
15222 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15223 RAND_egd() and RAND_status(). In the command line application,
15224 the EGD socket can be specified like a seed file using RANDFILE
15229 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15230 Some CAs (e.g. Verisign) distribute certificates in this form.
15234 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15235 list to exclude them. This means that no special compilation option
15236 is needed to use anonymous DH: it just needs to be included in the
15241 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15242 EVP_MD_type. The old functionality is available in a new macro called
15243 EVP_MD_md(). Change code that uses it and update docs.
15247 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15248 where the `void *` argument is replaced by a function pointer argument.
15249 Previously `void *` was abused to point to functions, which works on
15250 many platforms, but is not correct. As these functions are usually
15251 called by macros defined in OpenSSL header files, most source code
15252 should work without changes.
15256 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15257 sections with information on -D... compiler switches used for
15258 compiling the library so that applications can see them. To enable
15259 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15260 must be defined. E.g.,
15261 #define OPENSSL_ALGORITHM_DEFINES
15262 #include <openssl/opensslconf.h>
15263 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15265 *Richard Levitte, Ulf and Bodo Möller*
15267 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15272 * Change the 'other' type in certificate aux info to a STACK_OF
15273 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15274 the required ASN1 format: arbitrary types determined by an OID.
15278 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15279 argument to 'req'. This is not because the function is newer or
15280 better than others it just uses the work 'NEW' in the certificate
15281 request header lines. Some software needs this.
15285 * Reorganise password command line arguments: now passwords can be
15286 obtained from various sources. Delete the PEM_cb function and make
15287 it the default behaviour: i.e. if the callback is NULL and the
15288 usrdata argument is not NULL interpret it as a null terminated pass
15289 phrase. If usrdata and the callback are NULL then the pass phrase
15290 is prompted for as usual.
15294 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15295 the support is automatically enabled. The resulting binaries will
15296 autodetect the card and use it if present.
15298 *Ben Laurie and Compaq Inc.*
15300 * Work around for Netscape hang bug. This sends certificate request
15301 and server done in one record. Since this is perfectly legal in the
15302 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15303 the bugs/SSLv3 entry for more info.
15307 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15311 * Add -rand argument to smime and pkcs12 applications and read/write
15316 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15320 * Add command line password options to the remaining applications.
15324 * Bug fix for BN_div_recp() for numerators with an even number of
15329 * More tests in bntest.c, and changed test_bn output.
15333 * ./config recognizes MacOS X now.
15337 * Bug fix for BN_div() when the first words of num and divisor are
15338 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15342 * Add support for various broken PKCS#8 formats, and command line
15343 options to produce them.
15347 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15348 get temporary BIGNUMs from a BN_CTX.
15352 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15357 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15358 include a #define from the old name to the new. The original intent
15359 was that statically linked binaries could for example just call
15360 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15361 link with digests. This never worked because SSLeay_add_all_digests()
15362 and SSLeay_add_all_ciphers() were in the same source file so calling
15363 one would link with the other. They are now in separate source files.
15367 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15371 * Use a less unusual form of the Miller-Rabin primality test (it used
15372 a binary algorithm for exponentiation integrated into the Miller-Rabin
15373 loop, our standard modexp algorithms are faster).
15377 * Support for the EBCDIC character set completed.
15379 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15381 * Source code cleanups: use const where appropriate, eliminate casts,
15382 use `void *` instead of `char *` in lhash.
15386 * Bugfix: ssl3_send_server_key_exchange was not restartable
15387 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15388 this the server could overwrite ephemeral keys that the client
15393 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15394 using 50 iterations of the Rabin-Miller test.
15396 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15397 iterations of the Rabin-Miller test as required by the appendix
15398 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15399 As BN_is_prime_fasttest includes trial division, DSA parameter
15400 generation becomes much faster.
15402 This implies a change for the callback functions in DSA_is_prime
15403 and DSA_generate_parameters: The callback function is called once
15404 for each positive witness in the Rabin-Miller test, not just
15405 occasionally in the inner loop; and the parameters to the
15406 callback function now provide an iteration count for the outer
15407 loop rather than for the current invocation of the inner loop.
15408 DSA_generate_parameters additionally can call the callback
15409 function with an 'iteration count' of -1, meaning that a
15410 candidate has passed the trial division test (when q is generated
15411 from an application-provided seed, trial division is skipped).
15415 * New function BN_is_prime_fasttest that optionally does trial
15416 division before starting the Rabin-Miller test and has
15417 an additional BN_CTX * argument (whereas BN_is_prime always
15418 has to allocate at least one BN_CTX).
15419 'callback(1, -1, cb_arg)' is called when a number has passed the
15420 trial division stage.
15424 * Fix for bug in CRL encoding. The validity dates weren't being handled
15429 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15433 * New function BN_pseudo_rand().
15437 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15438 bignum version of BN_from_montgomery() with the working code from
15439 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15444 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15445 made it impossible to use the same SSL_SESSION data structure in
15446 SSL2 clients in multiple threads.
15450 * The return value of RAND_load_file() no longer counts bytes obtained
15451 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15452 to seed the PRNG (previously an explicit byte count was required).
15454 *Ulf Möller, Bodo Möller*
15456 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15457 used `char *` instead of `void *` and had casts all over the place.
15461 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15465 * Retain source code compatibility for BN_prime_checks macro:
15466 BN_is_prime(..., BN_prime_checks, ...) now uses
15467 BN_prime_checks_for_size to determine the appropriate number of
15468 Rabin-Miller iterations.
15472 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15473 DH_CHECK_P_NOT_SAFE_PRIME.
15474 (Check if this is true? OpenPGP calls them "strong".)
15478 * Merge the functionality of "dh" and "gendh" programs into a new program
15479 "dhparam". The old programs are retained for now but will handle DH keys
15480 (instead of parameters) in future.
15484 * Make the ciphers, s_server and s_client programs check the return values
15485 when a new cipher list is set.
15489 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15490 ciphers. Before when the 56bit ciphers were enabled the sorting was
15493 The syntax for the cipher sorting has been extended to support sorting by
15494 cipher-strength (using the strength_bits hard coded in the tables).
15495 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15497 Fix a bug in the cipher-command parser: when supplying a cipher command
15498 string with an "undefined" symbol (neither command nor alphanumeric
15499 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15500 an error is flagged.
15502 Due to the strength-sorting extension, the code of the
15503 ssl_create_cipher_list() function was completely rearranged. I hope that
15504 the readability was also increased :-)
15506 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15508 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15509 for the first serial number and places 2 in the serial number file. This
15510 avoids problems when the root CA is created with serial number zero and
15511 the first user certificate has the same issuer name and serial number
15516 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15517 the new code. Add documentation for this stuff.
15521 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15522 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15523 structures and behave in an analogous way to the X509v3 functions:
15524 they shouldn't be called directly but wrapper functions should be used
15527 So we also now have some wrapper functions that call the X509at functions
15528 when passed certificate requests. (TO DO: similar things can be done with
15529 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15530 things. Some of these need some d2i or i2d and print functionality
15531 because they handle more complex structures.)
15535 * Add missing #ifndefs that caused missing symbols when building libssl
15536 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15537 NO_RSA in `ssl/s2*.c`.
15539 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15541 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15542 has a return value which indicates the quality of the random data
15543 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15544 error queue. New function RAND_pseudo_bytes() generates output that is
15545 guaranteed to be unique but not unpredictable. RAND_add is like
15546 RAND_seed, but takes an extra argument for an entropy estimate
15547 (RAND_seed always assumes full entropy).
15551 * Do more iterations of Rabin-Miller probable prime test (specifically,
15552 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15553 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15554 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15555 false-positive rate of at most 2^-80 for random input.
15559 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15563 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15564 in the 0.9.5 release), this returns the chain
15565 from an X509_CTX structure with a dup of the stack and all
15566 the X509 reference counts upped: so the stack will exist
15567 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15570 Also make SSL_SESSION_print() print out the verify return
15575 * Add manpage for the pkcs12 command. Also change the default
15576 behaviour so MAC iteration counts are used unless the new
15577 -nomaciter option is used. This improves file security and
15578 only older versions of MSIE (4.0 for example) need it.
15582 * Honor the no-xxx Configure options when creating .DEF files.
15586 * Add PKCS#10 attributes to field table: challengePassword,
15587 unstructuredName and unstructuredAddress. These are taken from
15588 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15589 international characters are used.
15591 More changes to X509_ATTRIBUTE code: allow the setting of types
15592 based on strings. Remove the 'loc' parameter when adding
15593 attributes because these will be a SET OF encoding which is sorted
15598 * Initial changes to the 'req' utility to allow request generation
15599 automation. This will allow an application to just generate a template
15600 file containing all the field values and have req construct the
15603 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15604 used all over the place including certificate requests and PKCS#7
15605 structures. They are currently handled manually where necessary with
15606 some primitive wrappers for PKCS#7. The new functions behave in a
15607 manner analogous to the X509 extension functions: they allow
15608 attributes to be looked up by NID and added.
15610 Later something similar to the X509V3 code would be desirable to
15611 automatically handle the encoding, decoding and printing of the
15612 more complex types. The string types like challengePassword can
15613 be handled by the string table functions.
15615 Also modified the multi byte string table handling. Now there is
15616 a 'global mask' which masks out certain types. The table itself
15617 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15618 is useful when for example there is only one permissible type
15619 (as in countryName) and using the mask might result in no valid
15624 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15625 SSL_get_peer_finished to allow applications to obtain the latest
15626 Finished messages sent to the peer or expected from the peer,
15627 respectively. (SSL_get_peer_finished is usually the Finished message
15628 actually received from the peer, otherwise the protocol will be aborted.)
15630 As the Finished message are message digests of the complete handshake
15631 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15632 be used for external authentication procedures when the authentication
15633 provided by SSL/TLS is not desired or is not enough.
15637 * Enhanced support for Alpha Linux is added. Now ./config checks if
15638 the host supports BWX extension and if Compaq C is present on the
15639 $PATH. Just exploiting of the BWX extension results in 20-30%
15640 performance kick for some algorithms, e.g. DES and RC4 to mention
15641 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15646 * Add support for MS "fast SGC". This is arguably a violation of the
15647 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15648 weak crypto and after checking the certificate is SGC a second one
15649 with strong crypto. MS SGC stops the first handshake after receiving
15650 the server certificate message and sends a second client hello. Since
15651 a server will typically do all the time consuming operations before
15652 expecting any further messages from the client (server key exchange
15653 is the most expensive) there is little difference between the two.
15655 To get OpenSSL to support MS SGC we have to permit a second client
15656 hello message after we have sent server done. In addition we have to
15657 reset the MAC if we do get this second client hello.
15661 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15662 if a DER encoded private key is RSA or DSA traditional format. Changed
15663 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15664 format DER encoded private key. Newer code should use PKCS#8 format which
15665 has the key type encoded in the ASN1 structure. Added DER private key
15666 support to pkcs8 application.
15670 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15671 ciphersuites has been selected (as required by the SSL 3/TLS 1
15672 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15673 is set, we interpret this as a request to violate the specification
15674 (the worst that can happen is a handshake failure, and 'correct'
15675 behaviour would result in a handshake failure anyway).
15679 * In SSL_CTX_add_session, take into account that there might be multiple
15680 SSL_SESSION structures with the same session ID (e.g. when two threads
15681 concurrently obtain them from an external cache).
15682 The internal cache can handle only one SSL_SESSION with a given ID,
15683 so if there's a conflict, we now throw out the old one to achieve
15688 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15689 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15690 some routines that use cipher OIDs: some ciphers do not have OIDs
15691 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15696 * Simplify the trust setting structure and code. Now we just have
15697 two sequences of OIDs for trusted and rejected settings. These will
15698 typically have values the same as the extended key usage extension
15699 and any application specific purposes.
15701 The trust checking code now has a default behaviour: it will just
15702 check for an object with the same NID as the passed id. Functions can
15703 be provided to override either the default behaviour or the behaviour
15704 for a given id. SSL client, server and email already have functions
15705 in place for compatibility: they check the NID and also return "trusted"
15706 if the certificate is self signed.
15710 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15711 traditional format into an EVP_PKEY structure.
15715 * Add a password callback function PEM_cb() which either prompts for
15716 a password if usr_data is NULL or otherwise assumes it is a null
15717 terminated password. Allow passwords to be passed on command line
15718 environment or config files in a few more utilities.
15722 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15723 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15724 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15725 Update documentation.
15729 * Support for ASN1 "NULL" type. This could be handled before by using
15730 ASN1_TYPE but there wasn't any function that would try to read a NULL
15731 and produce an error if it couldn't. For compatibility we also have
15732 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15733 don't allocate anything because they don't need to.
15737 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15740 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15742 * Rebuild of the memory allocation routines used by OpenSSL code and
15743 possibly others as well. The purpose is to make an interface that
15744 provide hooks so anyone can build a separate set of allocation and
15745 deallocation routines to be used by OpenSSL, for example memory
15746 pool implementations, or something else, which was previously hard
15747 since Malloc(), Realloc() and Free() were defined as macros having
15748 the values malloc, realloc and free, respectively (except for Win32
15749 compilations). The same is provided for memory debugging code.
15750 OpenSSL already comes with functionality to find memory leaks, but
15751 this gives people a chance to debug other memory problems.
15753 With these changes, a new set of functions and macros have appeared:
15755 CRYPTO_set_mem_debug_functions() [F]
15756 CRYPTO_get_mem_debug_functions() [F]
15757 CRYPTO_dbg_set_options() [F]
15758 CRYPTO_dbg_get_options() [F]
15759 CRYPTO_malloc_debug_init() [M]
15761 The memory debug functions are NULL by default, unless the library
15762 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15763 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15764 gives the standard debugging functions that come with OpenSSL) or
15765 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15766 provided by the library user) must be used. When the standard
15767 debugging functions are used, CRYPTO_dbg_set_options can be used to
15768 request additional information:
15769 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15770 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15772 Also, things like CRYPTO_set_mem_functions will always give the
15773 expected result (the new set of functions is used for allocation
15774 and deallocation) at all times, regardless of platform and compiler
15777 To finish it up, some functions that were never use in any other
15778 way than through macros have a new API and new semantic:
15780 CRYPTO_dbg_malloc()
15781 CRYPTO_dbg_realloc()
15784 All macros of value have retained their old syntax.
15786 *Richard Levitte and Bodo Moeller*
15788 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15789 ordering of SMIMECapabilities wasn't in "strength order" and there
15790 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15795 * Some ASN1 types with illegal zero length encoding (INTEGER,
15796 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15798 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15800 * Merge in my S/MIME library for OpenSSL. This provides a simple
15801 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15802 functionality to handle multipart/signed properly) and a utility
15803 called 'smime' to call all this stuff. This is based on code I
15804 originally wrote for Celo who have kindly allowed it to be
15805 included in OpenSSL.
15809 * Add variants des_set_key_checked and des_set_key_unchecked of
15810 des_set_key (aka des_key_sched). Global variable des_check_key
15811 decides which of these is called by des_set_key; this way
15812 des_check_key behaves as it always did, but applications and
15813 the library itself, which was buggy for des_check_key == 1,
15814 have a cleaner way to pick the version they need.
15818 * New function PKCS12_newpass() which changes the password of a
15823 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15824 dynamic mix. In both cases the ids can be used as an index into the
15825 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
15826 functions so they accept a list of the field values and the
15827 application doesn't need to directly manipulate the X509_TRUST
15832 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
15837 * Modify the way the V3 extension code looks up extensions. This now
15838 works in a similar way to the object code: we have some "standard"
15839 extensions in a static table which is searched with OBJ_bsearch()
15840 and the application can add dynamic ones if needed. The file
15841 crypto/x509v3/ext_dat.h now has the info: this file needs to be
15842 updated whenever a new extension is added to the core code and kept
15843 in ext_nid order. There is a simple program 'tabtest.c' which checks
15844 this. New extensions are not added too often so this file can readily
15845 be maintained manually.
15847 There are two big advantages in doing things this way. The extensions
15848 can be looked up immediately and no longer need to be "added" using
15849 X509V3_add_standard_extensions(): this function now does nothing.
15850 Side note: I get *lots* of email saying the extension code doesn't
15851 work because people forget to call this function.
15852 Also no dynamic allocation is done unless new extensions are added:
15853 so if we don't add custom extensions there is no need to call
15854 X509V3_EXT_cleanup().
15858 * Modify enc utility's salting as follows: make salting the default. Add a
15859 magic header, so unsalted files fail gracefully instead of just decrypting
15860 to garbage. This is because not salting is a big security hole, so people
15861 should be discouraged from doing it.
15865 * Fixes and enhancements to the 'x509' utility. It allowed a message
15866 digest to be passed on the command line but it only used this
15867 parameter when signing a certificate. Modified so all relevant
15868 operations are affected by the digest parameter including the
15869 -fingerprint and -x509toreq options. Also -x509toreq choked if a
15870 DSA key was used because it didn't fix the digest.
15874 * Initial certificate chain verify code. Currently tests the untrusted
15875 certificates for consistency with the verify purpose (which is set
15876 when the X509_STORE_CTX structure is set up) and checks the pathlength.
15878 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
15879 this is because it will reject chains with invalid extensions whereas
15880 every previous version of OpenSSL and SSLeay made no checks at all.
15882 Trust code: checks the root CA for the relevant trust settings. Trust
15883 settings have an initial value consistent with the verify purpose: e.g.
15884 if the verify purpose is for SSL client use it expects the CA to be
15885 trusted for SSL client use. However the default value can be changed to
15886 permit custom trust settings: one example of this would be to only trust
15887 certificates from a specific "secure" set of CAs.
15889 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
15890 which should be used for version portability: especially since the
15891 verify structure is likely to change more often now.
15893 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
15894 to set them. If not set then assume SSL clients will verify SSL servers
15897 Two new options to the verify program: -untrusted allows a set of
15898 untrusted certificates to be passed in and -purpose which sets the
15899 intended purpose of the certificate. If a purpose is set then the
15900 new chain verify code is used to check extension consistency.
15904 * Support for the authority information access extension.
15908 * Modify RSA and DSA PEM read routines to transparently handle
15909 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
15910 public keys in a format compatible with certificate
15911 SubjectPublicKeyInfo structures. Unfortunately there were already
15912 functions called *_PublicKey_* which used various odd formats so
15913 these are retained for compatibility: however the DSA variants were
15914 never in a public release so they have been deleted. Changed dsa/rsa
15915 utilities to handle the new format: note no releases ever handled public
15916 keys so we should be OK.
15918 The primary motivation for this change is to avoid the same fiasco
15919 that dogs private keys: there are several incompatible private key
15920 formats some of which are standard and some OpenSSL specific and
15921 require various evil hacks to allow partial transparent handling and
15922 even then it doesn't work with DER formats. Given the option anything
15923 other than PKCS#8 should be dumped: but the other formats have to
15924 stay in the name of compatibility.
15926 With public keys and the benefit of hindsight one standard format
15927 is used which works with EVP_PKEY, RSA or DSA structures: though
15928 it clearly returns an error if you try to read the wrong kind of key.
15930 Added a -pubkey option to the 'x509' utility to output the public key.
15931 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
15932 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
15933 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
15934 that do the same as the `EVP_PKEY_assign_*()` except they up the
15935 reference count of the added key (they don't "swallow" the
15940 * Fixes to crypto/x509/by_file.c the code to read in certificates and
15941 CRLs would fail if the file contained no certificates or no CRLs:
15942 added a new function to read in both types and return the number
15943 read: this means that if none are read it will be an error. The
15944 DER versions of the certificate and CRL reader would always fail
15945 because it isn't possible to mix certificates and CRLs in DER format
15946 without choking one or the other routine. Changed this to just read
15947 a certificate: this is the best we can do. Also modified the code
15948 in apps/verify.c to take notice of return codes: it was previously
15949 attempting to read in certificates from NULL pointers and ignoring
15950 any errors: this is one reason why the cert and CRL reader seemed
15951 to work. It doesn't check return codes from the default certificate
15952 routines: these may well fail if the certificates aren't installed.
15956 * Code to support otherName option in GeneralName.
15960 * First update to verify code. Change the verify utility
15961 so it warns if it is passed a self signed certificate:
15962 for consistency with the normal behaviour. X509_verify
15963 has been modified to it will now verify a self signed
15964 certificate if *exactly* the same certificate appears
15965 in the store: it was previously impossible to trust a
15966 single self signed certificate. This means that:
15967 openssl verify ss.pem
15968 now gives a warning about a self signed certificate but
15969 openssl verify -CAfile ss.pem ss.pem
15974 * For servers, store verify_result in SSL_SESSION data structure
15975 (and add it to external session representation).
15976 This is needed when client certificate verifications fails,
15977 but an application-provided verification callback (set by
15978 SSL_CTX_set_cert_verify_callback) allows accepting the session
15979 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
15980 but returns 1): When the session is reused, we have to set
15981 ssl->verify_result to the appropriate error code to avoid
15984 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
15986 * Fix a bug in the new PKCS#7 code: it didn't consider the
15987 case in PKCS7_dataInit() where the signed PKCS7 structure
15988 didn't contain any existing data because it was being created.
15990 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
15992 * Add a salt to the key derivation routines in enc.c. This
15993 forms the first 8 bytes of the encrypted file. Also add a
15994 -S option to allow a salt to be input on the command line.
15998 * New function X509_cmp(). Oddly enough there wasn't a function
15999 to compare two certificates. We do this by working out the SHA1
16000 hash and comparing that. X509_cmp() will be needed by the trust
16005 * SSL_get1_session() is like SSL_get_session(), but increments
16006 the reference count in the SSL_SESSION returned.
16008 *Geoff Thorpe <geoff@eu.c2.net>*
16010 * Fix for 'req': it was adding a null to request attributes.
16011 Also change the X509_LOOKUP and X509_INFO code to handle
16012 certificate auxiliary information.
16016 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16021 * Add the possibility to add extra information to the memory leak
16022 detecting output, to form tracebacks, showing from where each
16023 allocation was originated: CRYPTO_push_info("constant string") adds
16024 the string plus current file name and line number to a per-thread
16025 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16026 is like calling CYRPTO_pop_info() until the stack is empty.
16027 Also updated memory leak detection code to be multi-thread-safe.
16031 * Add options -text and -noout to pkcs7 utility and delete the
16032 encryption options which never did anything. Update docs.
16036 * Add options to some of the utilities to allow the pass phrase
16037 to be included on either the command line (not recommended on
16038 OSes like Unix) or read from the environment. Update the
16039 manpages and fix a few bugs.
16043 * Add a few manpages for some of the openssl commands.
16047 * Fix the -revoke option in ca. It was freeing up memory twice,
16048 leaking and not finding already revoked certificates.
16052 * Extensive changes to support certificate auxiliary information.
16053 This involves the use of X509_CERT_AUX structure and X509_AUX
16054 functions. An X509_AUX function such as PEM_read_X509_AUX()
16055 can still read in a certificate file in the usual way but it
16056 will also read in any additional "auxiliary information". By
16057 doing things this way a fair degree of compatibility can be
16058 retained: existing certificates can have this information added
16059 using the new 'x509' options.
16061 Current auxiliary information includes an "alias" and some trust
16062 settings. The trust settings will ultimately be used in enhanced
16063 certificate chain verification routines: currently a certificate
16064 can only be trusted if it is self signed and then it is trusted
16069 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16070 The problem was that one of the replacement routines had not been working
16071 since SSLeay releases. For now the offending routine has been replaced
16072 with non-optimised assembler. Even so, this now gives around 95%
16073 performance improvement for 1024 bit RSA signs.
16077 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16078 handling. Most clients have the effective key size in bits equal to
16079 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16080 A few however don't do this and instead use the size of the decrypted key
16081 to determine the RC2 key length and the AlgorithmIdentifier to determine
16082 the effective key length. In this case the effective key length can still
16083 be 40 bits but the key length can be 168 bits for example. This is fixed
16084 by manually forcing an RC2 key into the EVP_PKEY structure because the
16085 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16086 the key length and effective key length are equal.
16090 * Add a bunch of functions that should simplify the creation of
16091 X509_NAME structures. Now you should be able to do:
16092 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16093 and have it automatically work out the correct field type and fill in
16094 the structures. The more adventurous can try:
16095 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16096 and it will (hopefully) work out the correct multibyte encoding.
16100 * Change the 'req' utility to use the new field handling and multibyte
16101 copy routines. Before the DN field creation was handled in an ad hoc
16102 way in req, ca, and x509 which was rather broken and didn't support
16103 BMPStrings or UTF8Strings. Since some software doesn't implement
16104 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16105 using the dirstring_type option. See the new comment in the default
16106 openssl.cnf for more info.
16110 * Make crypto/rand/md_rand.c more robust:
16111 - Assure unique random numbers after fork().
16112 - Make sure that concurrent threads access the global counter and
16113 md serializably so that we never lose entropy in them
16114 or use exactly the same state in multiple threads.
16115 Access to the large state is not always serializable because
16116 the additional locking could be a performance killer, and
16117 md should be large enough anyway.
16121 * New file apps/app_rand.c with commonly needed functionality
16122 for handling the random seed file.
16124 Use the random seed file in some applications that previously did not:
16126 dsaparam -genkey (which also ignored its '-rand' option),
16129 x509 (when signing).
16130 Except on systems with /dev/urandom, it is crucial to have a random
16131 seed file at least for key creation, DSA signing, and for DH exchanges;
16132 for RSA signatures we could do without one.
16134 gendh and gendsa (unlike genrsa) used to read only the first byte
16135 of each file listed in the '-rand' option. The function as previously
16136 found in genrsa is now in app_rand.c and is used by all programs
16137 that support '-rand'.
16141 * In RAND_write_file, use mode 0600 for creating files;
16142 don't just chmod when it may be too late.
16146 * Report an error from X509_STORE_load_locations
16147 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16151 * New function ASN1_mbstring_copy() this copies a string in either
16152 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16153 into an ASN1_STRING type. A mask of permissible types is passed
16154 and it chooses the "minimal" type to use or an error if not type
16159 * Add function equivalents to the various macros in asn1.h. The old
16160 macros are retained with an `M_` prefix. Code inside the library can
16161 use the `M_` macros. External code (including the openssl utility)
16162 should *NOT* in order to be "shared library friendly".
16166 * Add various functions that can check a certificate's extensions
16167 to see if it usable for various purposes such as SSL client,
16168 server or S/MIME and CAs of these types. This is currently
16169 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16170 verification. Also added a -purpose flag to x509 utility to
16171 print out all the purposes.
16175 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16180 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16181 for, obtain and decode and extension and obtain its critical flag.
16182 This allows all the necessary extension code to be handled in a
16183 single function call.
16187 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16188 platforms. See crypto/rc4/rc4_enc.c for further details.
16192 * New -noout option to asn1parse. This causes no output to be produced
16193 its main use is when combined with -strparse and -out to extract data
16194 from a file (which may not be in ASN.1 format).
16198 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16199 when producing the local key id.
16201 *Richard Levitte <levitte@stacken.kth.se>*
16203 * New option -dhparam in s_server. This allows a DH parameter file to be
16204 stated explicitly. If it is not stated then it tries the first server
16205 certificate file. The previous behaviour hard coded the filename
16210 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16211 a public key to be input or output. For example:
16212 openssl rsa -in key.pem -pubout -out pubkey.pem
16213 Also added necessary DSA public key functions to handle this.
16217 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16218 in the message. This was handled by allowing
16219 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16221 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16223 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16224 to the end of the strings whereas this didn't. This would cause problems
16225 if strings read with d2i_ASN1_bytes() were later modified.
16227 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16229 * Fix for base64 decode bug. When a base64 bio reads only one line of
16230 data and it contains EOF it will end up returning an error. This is
16231 caused by input 46 bytes long. The cause is due to the way base64
16232 BIOs find the start of base64 encoded data. They do this by trying a
16233 trial decode on each line until they find one that works. When they
16234 do a flag is set and it starts again knowing it can pass all the
16235 data directly through the decoder. Unfortunately it doesn't reset
16236 the context it uses. This means that if EOF is reached an attempt
16237 is made to pass two EOFs through the context and this causes the
16238 resulting error. This can also cause other problems as well. As is
16239 usual with these problems it takes *ages* to find and the fix is
16240 trivial: move one line.
16242 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16244 * Ugly workaround to get s_client and s_server working under Windows. The
16245 old code wouldn't work because it needed to select() on sockets and the
16246 tty (for keypresses and to see if data could be written). Win32 only
16247 supports select() on sockets so we select() with a 1s timeout on the
16248 sockets and then see if any characters are waiting to be read, if none
16249 are present then we retry, we also assume we can always write data to
16250 the tty. This isn't nice because the code then blocks until we've
16251 received a complete line of data and it is effectively polling the
16252 keyboard at 1s intervals: however it's quite a bit better than not
16253 working at all :-) A dedicated Windows application might handle this
16254 with an event loop for example.
16258 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16259 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16260 will be called when RSA_sign() and RSA_verify() are used. This is useful
16261 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16262 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16263 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16264 This necessitated the support of an extra signature type NID_md5_sha1
16265 for SSL signatures and modifications to the SSL library to use it instead
16266 of calling RSA_public_decrypt() and RSA_private_encrypt().
16270 * Add new -verify -CAfile and -CApath options to the crl program, these
16271 will lookup a CRL issuers certificate and verify the signature in a
16272 similar way to the verify program. Tidy up the crl program so it
16273 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16274 less strict. It will now permit CRL extensions even if it is not
16275 a V2 CRL: this will allow it to tolerate some broken CRLs.
16279 * Initialize all non-automatic variables each time one of the openssl
16280 sub-programs is started (this is necessary as they may be started
16281 multiple times from the "OpenSSL>" prompt).
16283 *Lennart Bang, Bodo Moeller*
16285 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16286 removing all other RSA functionality (this is what NO_RSA does). This
16287 is so (for example) those in the US can disable those operations covered
16288 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16293 * Non-copying interface to BIO pairs.
16294 (still largely untested)
16298 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16299 ASCII string. This was handled independently in various places before.
16303 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16304 UTF8 strings a character at a time.
16308 * Use client_version from client hello to select the protocol
16309 (s23_srvr.c) and for RSA client key exchange verification
16310 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16314 * Add various utility functions to handle SPKACs, these were previously
16315 handled by poking round in the structure internals. Added new function
16316 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16317 print, verify and generate SPKACs. Based on an original idea from
16318 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16322 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16326 * Allow the config file extension section to be overwritten on the
16327 command line. Based on an original idea from Massimiliano Pala
16328 <madwolf@comune.modena.it>. The new option is called -extensions
16329 and can be applied to ca, req and x509. Also -reqexts to override
16330 the request extensions in req and -crlexts to override the crl extensions
16335 * Add new feature to the SPKAC handling in ca. Now you can include
16336 the same field multiple times by preceding it by "XXXX." for example:
16339 this is the same syntax as used in the req config file.
16343 * Allow certificate extensions to be added to certificate requests. These
16344 are specified in a 'req_extensions' option of the req section of the
16345 config file. They can be printed out with the -text option to req but
16346 are otherwise ignored at present.
16350 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16351 data read consists of only the final block it would not decrypted because
16352 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16353 A misplaced 'break' also meant the decrypted final block might not be
16354 copied until the next read.
16358 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16359 a few extra parameters to the DH structure: these will be useful if
16360 for example we want the value of 'q' or implement X9.42 DH.
16364 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16365 provides hooks that allow the default DSA functions or functions on a
16366 "per key" basis to be replaced. This allows hardware acceleration and
16367 hardware key storage to be handled without major modification to the
16368 library. Also added low level modexp hooks and CRYPTO_EX structure and
16369 associated functions.
16373 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16374 as "read only": it can't be written to and the buffer it points to will
16375 not be freed. Reading from a read only BIO is much more efficient than
16376 a normal memory BIO. This was added because there are several times when
16377 an area of memory needs to be read from a BIO. The previous method was
16378 to create a memory BIO and write the data to it, this results in two
16379 copies of the data and an O(n^2) reading algorithm. There is a new
16380 function BIO_new_mem_buf() which creates a read only memory BIO from
16381 an area of memory. Also modified the PKCS#7 routines to use read only
16386 * Bugfix: ssl23_get_client_hello did not work properly when called in
16387 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16388 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16389 but a retry condition occurred while trying to read the rest.
16393 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16394 NID_pkcs7_encrypted by default: this was wrong since this should almost
16395 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16396 the encrypted data type: this is a more sensible place to put it and it
16397 allows the PKCS#12 code to be tidied up that duplicated this
16402 * Changed obj_dat.pl script so it takes its input and output files on
16403 the command line. This should avoid shell escape redirection problems
16408 * Initial support for certificate extension requests, these are included
16409 in things like Xenroll certificate requests. Included functions to allow
16410 extensions to be obtained and added.
16414 * -crlf option to s_client and s_server for sending newlines as
16415 CRLF (as required by many protocols).
16419 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16421 * Install libRSAglue.a when OpenSSL is built with RSAref.
16423 *Ralf S. Engelschall*
16425 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16427 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16429 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16434 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16435 DH parameters/keys (q is lost during that conversion, but the resulting
16436 DH parameters contain its length).
16438 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16439 much faster than DH_generate_parameters (which creates parameters
16440 where `p = 2*q + 1`), and also the smaller q makes DH computations
16441 much more efficient (160-bit exponentiation instead of 1024-bit
16442 exponentiation); so this provides a convenient way to support DHE
16443 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16444 utter importance to use
16445 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16447 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16448 when such DH parameters are used, because otherwise small subgroup
16449 attacks may become possible!
16453 * Avoid memory leak in i2d_DHparams.
16457 * Allow the -k option to be used more than once in the enc program:
16458 this allows the same encrypted message to be read by multiple recipients.
16462 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16463 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16464 it will always use the numerical form of the OID, even if it has a short
16469 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16470 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16471 otherwise bn_mod_exp was called. In the case of hardware keys for example
16472 no private key components need be present and it might store extra data
16473 in the RSA structure, which cannot be accessed from bn_mod_exp.
16474 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16475 private key operations.
16479 * Added support for SPARC Linux.
16483 * pem_password_cb function type incompatibly changed from
16484 typedef int pem_password_cb(char *buf, int size, int rwflag);
16486 ....(char *buf, int size, int rwflag, void *userdata);
16487 so that applications can pass data to their callbacks:
16488 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16489 additional void * argument, which is just handed through whenever
16490 the password callback is called.
16492 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16494 New function SSL_CTX_set_default_passwd_cb_userdata.
16496 Compatibility note: As many C implementations push function arguments
16497 onto the stack in reverse order, the new library version is likely to
16498 interoperate with programs that have been compiled with the old
16499 pem_password_cb definition (PEM_whatever takes some data that
16500 happens to be on the stack as its last argument, and the callback
16501 just ignores this garbage); but there is no guarantee whatsoever that
16504 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16505 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16506 problems not only on Windows, but also on some Unix platforms.
16507 To avoid problematic command lines, these definitions are now in an
16508 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16509 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16513 * MIPS III/IV assembler module is reimplemented.
16517 * More DES library cleanups: remove references to srand/rand and
16518 delete an unused file.
16522 * Add support for the free Netwide assembler (NASM) under Win32,
16523 since not many people have MASM (ml) and it can be hard to obtain.
16524 This is currently experimental but it seems to work OK and pass all
16525 the tests. Check out INSTALL.W32 for info.
16529 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16530 without temporary keys kept an extra copy of the server key,
16531 and connections with temporary keys did not free everything in case
16536 * New function RSA_check_key and new openssl rsa option -check
16537 for verifying the consistency of RSA keys.
16539 *Ulf Moeller, Bodo Moeller*
16541 * Various changes to make Win32 compile work:
16542 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16543 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16544 comparison" warnings.
16545 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16549 * Add a debugging option to PKCS#5 v2 key generation function: when
16550 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16551 derived keys are printed to stderr.
16555 * Copy the flags in ASN1_STRING_dup().
16557 *Roman E. Pavlov <pre@mo.msk.ru>*
16559 * The x509 application mishandled signing requests containing DSA
16560 keys when the signing key was also DSA and the parameters didn't match.
16562 It was supposed to omit the parameters when they matched the signing key:
16563 the verifying software was then supposed to automatically use the CA's
16564 parameters if they were absent from the end user certificate.
16566 Omitting parameters is no longer recommended. The test was also
16567 the wrong way round! This was probably due to unusual behaviour in
16568 EVP_cmp_parameters() which returns 1 if the parameters match.
16569 This meant that parameters were omitted when they *didn't* match and
16570 the certificate was useless. Certificates signed with 'ca' didn't have
16573 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16575 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16576 The interface is as follows:
16577 Applications can use
16578 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16579 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16580 "off" is now the default.
16581 The library internally uses
16582 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16583 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16584 to disable memory-checking temporarily.
16586 Some inconsistent states that previously were possible (and were
16587 even the default) are now avoided.
16589 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16590 with each memory chunk allocated; this is occasionally more helpful
16591 than just having a counter.
16593 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16595 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16600 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16601 which largely parallels "options", but is for changing API behaviour,
16602 whereas "options" are about protocol behaviour.
16603 Initial "mode" flags are:
16605 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16606 a single record has been written.
16607 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16608 retries use the same buffer location.
16609 (But all of the contents must be
16614 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16617 * Fix problems with no-hmac etc.
16619 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16621 * New functions RSA_get_default_method(), RSA_set_method() and
16622 RSA_get_method(). These allows replacement of RSA_METHODs without having
16623 to mess around with the internals of an RSA structure.
16627 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16628 Also really enable memory leak checks in openssl.c and in some
16631 *Chad C. Mulligan, Bodo Moeller*
16633 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16634 up the length of negative integers. This has now been simplified to just
16635 store the length when it is first determined and use it later, rather
16636 than trying to keep track of where data is copied and updating it to
16638 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16640 * Add a new function PKCS7_signatureVerify. This allows the verification
16641 of a PKCS#7 signature but with the signing certificate passed to the
16642 function itself. This contrasts with PKCS7_dataVerify which assumes the
16643 certificate is present in the PKCS#7 structure. This isn't always the
16644 case: certificates can be omitted from a PKCS#7 structure and be
16645 distributed by "out of band" means (such as a certificate database).
16649 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16650 function prototypes in pem.h, also change util/mkdef.pl to add the
16651 necessary function names.
16655 * mk1mf.pl (used by Windows builds) did not properly read the
16656 options set by Configure in the top level Makefile, and Configure
16657 was not even able to write more than one option correctly.
16658 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16662 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16663 file to be loaded from a BIO or FILE pointer. The BIO version will
16664 for example allow memory BIOs to contain config info.
16668 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16669 Whoever hopes to achieve shared-library compatibility across versions
16670 must use this, not the compile-time macro.
16671 (Exercise 0.9.4: Which is the minimum library version required by
16673 Note: All this applies only to multi-threaded programs, others don't
16678 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16679 through a BIO pair triggered the default case, i.e.
16680 SSLerr(...,SSL_R_UNKNOWN_STATE).
16684 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16685 can use the SSL library even if none of the specific BIOs is
16690 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16691 for the encoded length.
16693 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16695 * Add initial documentation of the X509V3 functions.
16699 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16700 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16701 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16702 secure PKCS#8 private key format with a high iteration count.
16706 * Fix determination of Perl interpreter: A perl or perl5
16707 *directory* in $PATH was also accepted as the interpreter.
16709 *Ralf S. Engelschall*
16711 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16712 wrong with it but it was very old and did things like calling
16713 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16714 unusual formatting.
16718 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16719 to use the new extension code.
16723 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16724 with macros. This should make it easier to change their form, add extra
16725 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16730 * Add to configuration table a new entry that can specify an alternative
16731 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16732 according to Mark Crispin <MRC@Panda.COM>.
16736 * DES CBC did not update the IV. Weird.
16740 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16741 Changing the behaviour of the former might break existing programs --
16742 where IV updating is needed, des_ncbc_encrypt can be used.
16745 * When bntest is run from "make test" it drives bc to check its
16746 calculations, as well as internally checking them. If an internal check
16747 fails, it needs to cause bc to give a non-zero result or make test carries
16748 on without noticing the failure. Fixed.
16752 * DES library cleanups.
16756 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16757 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16758 ciphers. NOTE: although the key derivation function has been verified
16759 against some published test vectors it has not been extensively tested
16760 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16765 * Instead of "mkdir -p", which is not fully portable, use new
16766 Perl script "util/mkdir-p.pl".
16770 * Rewrite the way password based encryption (PBE) is handled. It used to
16771 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16772 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16773 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16774 the 'parameter' field of the AlgorithmIdentifier is passed to the
16775 underlying key generation function so it must do its own ASN1 parsing.
16776 This has also changed the EVP_PBE_CipherInit() function which now has a
16777 'parameter' argument instead of literal salt and iteration count values
16778 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16782 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16783 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16784 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16785 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16786 value was just used as a "magic string" and not used directly its
16787 value doesn't matter.
16791 * Introduce some semblance of const correctness to BN. Shame C doesn't
16796 * "linux-sparc64" configuration (ultrapenguin).
16798 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16799 "linux-sparc" configuration.
16801 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16803 * config now generates no-xxx options for missing ciphers.
16807 * Support the EBCDIC character set (work in progress).
16808 File ebcdic.c not yet included because it has a different license.
16810 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16812 * Support BS2000/OSD-POSIX.
16814 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16816 * Make callbacks for key generation use `void *` instead of `char *`.
16820 * Make S/MIME samples compile (not yet tested).
16824 * Additional typesafe stacks.
16828 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
16832 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
16834 * New configuration variant "sco5-gcc".
16836 * Updated some demos.
16838 *Sean O Riordain, Wade Scholine*
16840 * Add missing BIO_free at exit of pkcs12 application.
16844 * Fix memory leak in conf.c.
16848 * Updates for Win32 to assembler version of MD5.
16852 * Set #! path to perl in apps/der_chop to where we found it
16853 instead of using a fixed path.
16857 * SHA library changes for irix64-mips4-cc.
16861 * Improvements for VMS support.
16865 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
16867 * Bignum library bug fix. IRIX 6 passes "make test" now!
16868 This also avoids the problems with SC4.2 and unpatched SC5.
16870 *Andy Polyakov <appro@fy.chalmers.se>*
16872 * New functions sk_num, sk_value and sk_set to replace the previous macros.
16873 These are required because of the typesafe stack would otherwise break
16874 existing code. If old code used a structure member which used to be STACK
16875 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
16876 sk_num or sk_value it would produce an error because the num, data members
16877 are not present in STACK_OF. Now it just produces a warning. sk_set
16878 replaces the old method of assigning a value to sk_value
16879 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
16880 that does this will no longer work (and should use sk_set instead) but
16881 this could be regarded as a "questionable" behaviour anyway.
16885 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
16886 correctly handle encrypted S/MIME data.
16890 * Change type of various DES function arguments from des_cblock
16891 (which means, in function argument declarations, pointer to char)
16892 to des_cblock * (meaning pointer to array with 8 char elements),
16893 which allows the compiler to do more typechecking; it was like
16894 that back in SSLeay, but with lots of ugly casts.
16896 Introduce new type const_des_cblock.
16900 * Reorganise the PKCS#7 library and get rid of some of the more obvious
16901 problems: find RecipientInfo structure that matches recipient certificate
16902 and initialise the ASN1 structures properly based on passed cipher.
16906 * Belatedly make the BN tests actually check the results.
16910 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
16911 to and from BNs: it was completely broken. New compilation option
16912 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
16913 key elements as negative integers.
16917 * Reorganize and speed up MD5.
16919 *Andy Polyakov <appro@fy.chalmers.se>*
16923 *Richard Levitte <richard@levitte.org>*
16925 * New option -out to asn1parse to allow the parsed structure to be
16926 output to a file. This is most useful when combined with the -strparse
16927 option to examine the output of things like OCTET STRINGS.
16931 * Make SSL library a little more fool-proof by not requiring any longer
16932 that `SSL_set_{accept,connect}_state` be called before
16933 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
16934 in many applications because usually everything *appeared* to work as
16935 intended anyway -- now it really works as intended).
16939 * Move openssl.cnf out of lib/.
16943 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
16944 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
16945 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
16947 *Ralf S. Engelschall*
16949 * Various fixes to the EVP and PKCS#7 code. It may now be able to
16950 handle PKCS#7 enveloped data properly.
16952 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
16954 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
16955 copying pointers. The cert_st handling is changed by this in
16956 various ways (and thus what used to be known as ctx->default_cert
16957 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
16958 any longer when s->cert does not give us what we need).
16959 ssl_cert_instantiate becomes obsolete by this change.
16960 As soon as we've got the new code right (possibly it already is?),
16961 we have solved a couple of bugs of the earlier code where s->cert
16962 was used as if it could not have been shared with other SSL structures.
16964 Note that using the SSL API in certain dirty ways now will result
16965 in different behaviour than observed with earlier library versions:
16966 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
16967 does not influence s as it used to.
16969 In order to clean up things more thoroughly, inside SSL_SESSION
16970 we don't use CERT any longer, but a new structure SESS_CERT
16971 that holds per-session data (if available); currently, this is
16972 the peer's certificate chain and, for clients, the server's certificate
16973 and temporary key. CERT holds only those values that can have
16974 meaningful defaults in an SSL_CTX.
16978 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
16979 from the internal representation. Various PKCS#7 fixes: remove some
16980 evil casts and set the enc_dig_alg field properly based on the signing
16985 * Allow PKCS#12 password to be set from the command line or the
16986 environment. Let 'ca' get its config file name from the environment
16987 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
16992 * Allow certificate policies extension to use an IA5STRING for the
16993 organization field. This is contrary to the PKIX definition but
16994 VeriSign uses it and IE5 only recognises this form. Document 'x509'
16999 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17000 without disallowing inline assembler and the like for non-pedantic builds.
17004 * Support Borland C++ builder.
17006 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17012 * SHA-1 cleanups and performance enhancements.
17014 *Andy Polyakov <appro@fy.chalmers.se>*
17016 * Sparc v8plus assembler for the bignum library.
17018 *Andy Polyakov <appro@fy.chalmers.se>*
17020 * Accept any -xxx and +xxx compiler options in Configure.
17024 * Update HPUX configuration.
17028 * Add missing `sk_<type>_unshift()` function to safestack.h
17030 *Ralf S. Engelschall*
17032 * New function SSL_CTX_use_certificate_chain_file that sets the
17033 "extra_cert"s in addition to the certificate. (This makes sense
17034 only for "PEM" format files, as chains as a whole are not
17039 * Support verify_depth from the SSL API.
17040 x509_vfy.c had what can be considered an off-by-one-error:
17041 Its depth (which was not part of the external interface)
17042 was actually counting the number of certificates in a chain;
17043 now it really counts the depth.
17047 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17048 instead of X509err, which often resulted in confusing error
17049 messages since the error codes are not globally unique
17050 (e.g. an alleged error in ssl3_accept when a certificate
17051 didn't match the private key).
17053 * New function SSL_CTX_set_session_id_context that allows to set a default
17054 value (so that you don't need SSL_set_session_id_context for each
17055 connection using the SSL_CTX).
17059 * OAEP decoding bug fix.
17063 * Support INSTALL_PREFIX for package builders, as proposed by
17068 * New Configure options "threads" and "no-threads". For systems
17069 where the proper compiler options are known (currently Solaris
17070 and Linux), "threads" is the default.
17074 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17078 * Install various scripts to $(OPENSSLDIR)/misc, not to
17079 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17080 such as /usr/local/bin.
17084 * "make linux-shared" to build shared libraries.
17086 *Niels Poppe <niels@netbox.org>*
17088 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17092 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17093 extension adding in x509 utility.
17097 * Remove NOPROTO sections and error code comments.
17101 * Partial rewrite of the DEF file generator to now parse the ANSI
17106 * New Configure options --prefix=DIR and --openssldir=DIR.
17110 * Complete rewrite of the error code script(s). It is all now handled
17111 by one script at the top level which handles error code gathering,
17112 header rewriting and C source file generation. It should be much better
17113 than the old method: it now uses a modified version of Ulf's parser to
17114 read the ANSI prototypes in all header files (thus the old K&R definitions
17115 aren't needed for error creation any more) and do a better job of
17116 translating function codes into names. The old 'ASN1 error code embedded
17117 in a comment' is no longer necessary and it doesn't use .err files which
17118 have now been deleted. Also the error code call doesn't have to appear all
17119 on one line (which resulted in some large lines...).
17123 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17127 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17128 0 (which usually indicates a closed connection), but continue reading.
17132 * Fix some race conditions.
17136 * Add support for CRL distribution points extension. Add Certificate
17137 Policies and CRL distribution points documentation.
17141 * Move the autogenerated header file parts to crypto/opensslconf.h.
17145 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17146 8 of keying material. Merlin has also confirmed interop with this fix
17147 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17149 *Merlin Hughes <merlin@baltimore.ie>*
17151 * Fix lots of warnings.
17153 *Richard Levitte <levitte@stacken.kth.se>*
17155 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17156 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17158 *Richard Levitte <levitte@stacken.kth.se>*
17160 * Fix problems with sizeof(long) == 8.
17162 *Andy Polyakov <appro@fy.chalmers.se>*
17164 * Change functions to ANSI C.
17168 * Fix typos in error codes.
17170 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17172 * Remove defunct assembler files from Configure.
17176 * SPARC v8 assembler BIGNUM implementation.
17178 *Andy Polyakov <appro@fy.chalmers.se>*
17180 * Support for Certificate Policies extension: both print and set.
17181 Various additions to support the r2i method this uses.
17185 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17186 return a const string when you are expecting an allocated buffer.
17190 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17191 types DirectoryString and DisplayText.
17195 * Add code to allow r2i extensions to access the configuration database,
17196 add an LHASH database driver and add several ctx helper functions.
17200 * Fix an evil bug in bn_expand2() which caused various BN functions to
17201 fail when they extended the size of a BIGNUM.
17205 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17206 support typesafe stack.
17210 * Fix typo in SSL_[gs]et_options().
17212 *Nils Frostberg <nils@medcom.se>*
17214 * Delete various functions and files that belonged to the (now obsolete)
17215 old X509V3 handling code.
17219 * New Configure option "rsaref".
17223 * Don't auto-generate pem.h.
17227 * Introduce type-safe ASN.1 SETs.
17231 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17233 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17235 * Introduce type-safe STACKs. This will almost certainly break lots of code
17236 that links with OpenSSL (well at least cause lots of warnings), but fear
17237 not: the conversion is trivial, and it eliminates loads of evil casts. A
17238 few STACKed things have been converted already. Feel free to convert more.
17239 In the fullness of time, I'll do away with the STACK type altogether.
17243 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17244 specified in `<certfile>` by updating the entry in the index.txt file.
17245 This way one no longer has to edit the index.txt file manually for
17246 revoking a certificate. The -revoke option does the gory details now.
17248 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17250 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17251 `-text` option at all and this way the `-noout -text` combination was
17252 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17254 *Ralf S. Engelschall*
17256 * Make sure a corresponding plain text error message exists for the
17257 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17258 verify callback function determined that a certificate was revoked.
17260 *Ralf S. Engelschall*
17262 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17263 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17264 all available ciphers including rc5, which was forgotten until now.
17265 In order to let the testing shell script know which algorithms
17266 are available, a new (up to now undocumented) command
17267 `openssl list-cipher-commands` is used.
17271 * Bugfix: s_client occasionally would sleep in select() when
17272 it should have checked SSL_pending() first.
17276 * New functions DSA_do_sign and DSA_do_verify to provide access to
17277 the raw DSA values prior to ASN.1 encoding.
17281 * Tweaks to Configure
17283 *Niels Poppe <niels@netbox.org>*
17285 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17290 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17294 * New config option to avoid instructions that are illegal on the 80386.
17295 The default code is faster, but requires at least a 486.
17299 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17300 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17301 same as SSL2_VERSION anyway.
17305 * New "-showcerts" option for s_client.
17309 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17310 application. Various cleanups and fixes.
17314 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17315 modify error routines to work internally. Add error codes and PBE init
17316 to library startup routines.
17320 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17321 packing functions to asn1 and evp. Changed function names and error
17322 codes along the way.
17326 * PKCS12 integration: and so it begins... First of several patches to
17327 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17328 objects to objects.h
17332 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17333 and display support for Thawte strong extranet extension.
17337 * Add LinuxPPC support.
17339 *Jeff Dubrule <igor@pobox.org>*
17341 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17342 bn_div_words in alpha.s.
17344 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17346 * Make sure the RSA OAEP test is skipped under -DRSAref because
17347 OAEP isn't supported when OpenSSL is built with RSAref.
17349 *Ulf Moeller <ulf@fitug.de>*
17351 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17352 so they no longer are missing under -DNOPROTO.
17354 *Soren S. Jorvang <soren@t.dk>*
17356 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17358 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17359 doesn't work when the session is reused. Coming soon!
17363 * Fix a security hole, that allows sessions to be reused in the wrong
17364 context thus bypassing client cert protection! All software that uses
17365 client certs and session caches in multiple contexts NEEDS PATCHING to
17366 allow session reuse! A fuller solution is in the works.
17368 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17370 * Some more source tree cleanups (removed obsolete files
17371 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17372 permission on "config" script to be executable) and a fix for the INSTALL
17375 *Ulf Moeller <ulf@fitug.de>*
17377 * Remove some legacy and erroneous uses of malloc, free instead of
17380 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17382 * Make rsa_oaep_test return non-zero on error.
17384 *Ulf Moeller <ulf@fitug.de>*
17386 * Add support for native Solaris shared libraries. Configure
17387 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17388 if someone would make that last step automatic.
17390 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17392 * ctx_size was not built with the right compiler during "make links". Fixed.
17396 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17397 except NULL ciphers". This means the default cipher list will no longer
17398 enable NULL ciphers. They need to be specifically enabled e.g. with
17399 the string "DEFAULT:eNULL".
17403 * Fix to RSA private encryption routines: if p < q then it would
17404 occasionally produce an invalid result. This will only happen with
17405 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17409 * Be less restrictive and allow also `perl util/perlpath.pl
17410 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17411 because this way one can also use an interpreter named `perl5` (which is
17412 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17413 installed as `perl`).
17415 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17417 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17419 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17421 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17422 advapi32.lib to Win32 build and change the pem test comparison
17423 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17424 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17425 and crypto/des/ede_cbcm_enc.c.
17429 * DES quad checksum was broken on big-endian architectures. Fixed.
17433 * Comment out two functions in bio.h that aren't implemented. Fix up the
17434 Win32 test batch file so it (might) work again. The Win32 test batch file
17435 is horrible: I feel ill....
17439 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17440 in e_os.h. Audit of header files to check ANSI and non ANSI
17441 sections: 10 functions were absent from non ANSI section and not exported
17442 from Windows DLLs. Fixed up libeay.num for new functions.
17446 * Make `openssl version` output lines consistent.
17448 *Ralf S. Engelschall*
17450 * Fix Win32 symbol export lists for BIO functions: Added
17451 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17452 to ms/libeay{16,32}.def.
17454 *Ralf S. Engelschall*
17456 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17457 fine under Unix and passes some trivial tests I've now added. But the
17458 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17459 added to make sure no one expects that this stuff really works in the
17460 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17461 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17464 *Ralf S. Engelschall*
17466 * Fix the generation of two part addresses in perl.
17468 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17470 * Add config entry for Linux on MIPS.
17472 *John Tobey <jtobey@channel1.com>*
17474 * Make links whenever Configure is run, unless we are on Windoze.
17478 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17479 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17484 * Add a useful kludge to allow package maintainers to specify compiler and
17485 other platforms details on the command line without having to patch the
17486 Configure script every time: One now can use
17487 `perl Configure <id>:<details>`,
17488 i.e. platform ids are allowed to have details appended
17489 to them (separated by colons). This is treated as there would be a static
17490 pre-configured entry in Configure's %table under key `<id>` with value
17491 `<details>` and `perl Configure <id>` is called. So, when you want to
17492 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17493 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17494 now, which overrides the FreeBSD-elf entry on-the-fly.
17496 *Ralf S. Engelschall*
17498 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17502 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17503 on the `perl Configure ...` command line. This way one can compile
17504 OpenSSL libraries with Position Independent Code (PIC) which is needed
17505 for linking it into DSOs.
17507 *Ralf S. Engelschall*
17509 * Remarkably, export ciphers were totally broken and no-one had noticed!
17514 * Cleaned up the LICENSE document: The official contact for any license
17515 questions now is the OpenSSL core team under openssl-core@openssl.org.
17516 And add a paragraph about the dual-license situation to make sure people
17517 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17518 to the OpenSSL toolkit.
17520 *Ralf S. Engelschall*
17522 * General source tree makefile cleanups: Made `making xxx in yyy...`
17523 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17524 Additionally cleaned up the `make links` target: Remove unnecessary
17525 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17526 to speed processing and no longer clutter the display with confusing
17527 stuff. Instead only the actually done links are displayed.
17529 *Ralf S. Engelschall*
17531 * Permit null encryption ciphersuites, used for authentication only. It used
17532 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17533 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17538 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17539 signed attributes when verifying signatures (this would break them),
17540 the detached data encoding was wrong and public keys obtained using
17541 X509_get_pubkey() weren't freed.
17545 * Add text documentation for the BUFFER functions. Also added a work around
17546 to a Win95 console bug. This was triggered by the password read stuff: the
17547 last character typed gets carried over to the next fread(). If you were
17548 generating a new cert request using 'req' for example then the last
17549 character of the passphrase would be CR which would then enter the first
17554 * Added the new 'Includes OpenSSL Cryptography Software' button as
17555 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17556 button and can be used by applications based on OpenSSL to show the
17557 relationship to the OpenSSL project.
17559 *Ralf S. Engelschall*
17561 * Remove confusing variables in function signatures in files
17562 ssl/ssl_lib.c and ssl/ssl.h.
17564 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17566 * Don't install bss_file.c under PREFIX/include/
17568 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17570 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17571 functions that return function pointers and has support for NT specific
17572 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17573 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17574 unsigned to signed types: this was killing the Win32 compile.
17578 * Add new certificate file to stack functions,
17579 SSL_add_dir_cert_subjects_to_stack() and
17580 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17581 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17582 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17583 This means that Apache-SSL and similar packages don't have to mess around
17584 to add as many CAs as they want to the preferred list.
17588 * Experiment with doxygen documentation. Currently only partially applied to
17590 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17591 openssl.doxy as the configuration file.
17595 * Get rid of remaining C++-style comments which strict C compilers hate.
17597 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17599 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17600 compiled in by default: it has problems with large keys.
17604 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17605 DH private keys and/or callback functions which directly correspond to
17606 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17607 is needed for applications which have to configure certificates on a
17608 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17610 For the RSA certificate situation is makes no difference, but
17611 for the DSA certificate situation this fixes the "no shared cipher"
17612 problem where the OpenSSL cipher selection procedure failed because the
17613 temporary keys were not overtaken from the context and the API provided
17614 no way to reconfigure them.
17615 The new functions now let applications reconfigure the stuff and they
17616 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17617 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17618 non-public-API function ssl_cert_instantiate() is used as a helper
17619 function and also to reduce code redundancy inside ssl_rsa.c.
17621 *Ralf S. Engelschall*
17623 * Move s_server -dcert and -dkey options out of the undocumented feature
17624 area because they are useful for the DSA situation and should be
17625 recognized by the users.
17627 *Ralf S. Engelschall*
17629 * Fix the cipher decision scheme for export ciphers: the export bits are
17630 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17631 SSL_EXP_MASK. So, the original variable has to be used instead of the
17632 already masked variable.
17634 *Richard Levitte <levitte@stacken.kth.se>*
17636 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17638 *Richard Levitte <levitte@stacken.kth.se>*
17640 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17641 from `int` to `unsigned int` because it is a length and initialized by
17642 EVP_DigestFinal() which expects an `unsigned int *`.
17644 *Richard Levitte <levitte@stacken.kth.se>*
17646 * Don't hard-code path to Perl interpreter on shebang line of Configure
17647 script. Instead use the usual Shell->Perl transition trick.
17649 *Ralf S. Engelschall*
17651 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17652 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17653 -noout -modulus` as it's already the case for `openssl rsa -noout
17654 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17655 currently the public key is printed (a decision which was already done by
17656 `openssl dsa -modulus` in the past) which serves a similar purpose.
17657 Additionally the NO_RSA no longer completely removes the whole -modulus
17658 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17661 *Ralf S. Engelschall*
17663 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17664 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17666 *Arne Ansper <arne@ats.cyber.ee>*
17668 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17669 to be added. Now both 'req' and 'ca' can use new objects defined in the
17674 * Add cool BIO that does syslog (or event log on NT).
17676 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17678 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17679 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17680 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17681 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17685 * Add preliminary config info for new extension code.
17689 * Make RSA_NO_PADDING really use no padding.
17691 *Ulf Moeller <ulf@fitug.de>*
17693 * Generate errors when private/public key check is done.
17697 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17698 for some CRL extensions and new objects added.
17702 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17703 key usage extension and fuller support for authority key id.
17707 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17708 padding method for RSA, which is recommended for new applications in PKCS
17709 #1 v2.0 (RFC 2437, October 1998).
17710 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17711 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17712 against Bleichbacher's attack on RSA.
17713 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17716 * Updates to the new SSL compression code
17718 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17720 * Fix so that the version number in the master secret, when passed
17721 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17722 (because the server will not accept higher), that the version number
17723 is 0x03,0x01, not 0x03,0x00
17725 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17727 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17728 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17729 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17733 * Support for RAW extensions where an arbitrary extension can be
17734 created by including its DER encoding. See apps/openssl.cnf for
17739 * Make sure latest Perl versions don't interpret some generated C array
17740 code as Perl array code in the crypto/err/err_genc.pl script.
17742 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17744 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17745 not many people have the assembler. Various Win32 compilation fixes and
17746 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17747 build instructions.
17751 * Modify configure script 'Configure' to automatically create crypto/date.h
17752 file under Win32 and also build pem.h from pem.org. New script
17753 util/mkfiles.pl to create the MINFO file on environments that can't do a
17754 'make files': perl util/mkfiles.pl >MINFO should work.
17758 * Major rework of DES function declarations, in the pursuit of correctness
17759 and purity. As a result, many evil casts evaporated, and some weirdness,
17760 too. You may find this causes warnings in your code. Zapping your evil
17761 casts will probably fix them. Mostly.
17765 * Fix for a typo in asn1.h. Bug fix to object creation script
17766 obj_dat.pl. It considered a zero in an object definition to mean
17767 "end of object": none of the objects in objects.h have any zeros
17768 so it wasn't spotted.
17770 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17772 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17773 Masking (CBCM). In the absence of test vectors, the best I have been able
17774 to do is check that the decrypt undoes the encrypt, so far. Send me test
17775 vectors if you have them.
17779 * Correct calculation of key length for export ciphers (too much space was
17780 allocated for null ciphers). This has not been tested!
17784 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17785 message is now correct (it understands "crypto" and "ssl" on its
17786 command line). There is also now an "update" option. This will update
17787 the util/ssleay.num and util/libeay.num files with any new functions.
17789 perl util/mkdef.pl crypto ssl update
17790 it will update them.
17794 * Overhauled the Perl interface:
17795 - ported BN stuff to OpenSSL's different BN library
17796 - made the perl/ source tree CVS-aware
17797 - renamed the package from SSLeay to OpenSSL (the files still contain
17798 their history because I've copied them in the repository)
17799 - removed obsolete files (the test scripts will be replaced
17800 by better Test::Harness variants in the future)
17802 *Ralf S. Engelschall*
17804 * First cut for a very conservative source tree cleanup:
17805 1. merge various obsolete readme texts into doc/ssleay.txt
17806 where we collect the old documents and readme texts.
17807 2. remove the first part of files where I'm already sure that we no
17808 longer need them because of three reasons: either they are just temporary
17809 files which were left by Eric or they are preserved original files where
17810 I've verified that the diff is also available in the CVS via "cvs diff
17811 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17812 the crypto/md/ stuff).
17814 *Ralf S. Engelschall*
17816 * More extension code. Incomplete support for subject and issuer alt
17817 name, issuer and authority key id. Change the i2v function parameters
17818 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17819 what that's for :-) Fix to ASN1 macro which messed up
17820 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17824 * Preliminary support for ENUMERATED type. This is largely copied from the
17829 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
17831 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17833 * Make sure `make rehash` target really finds the `openssl` program.
17835 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17837 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
17838 like to hear about it if this slows down other processors.
17842 * Add CygWin32 platform information to Configure script.
17844 *Alan Batie <batie@aahz.jf.intel.com>*
17846 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
17848 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
17850 * New program nseq to manipulate netscape certificate sequences
17854 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
17859 * Fixes to BN code. Previously the default was to define BN_RECURSION
17860 but the BN code had some problems that would cause failures when
17861 doing certificate verification and some other functions.
17863 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17865 * Add ASN1 and PEM code to support netscape certificate sequences.
17869 * Add ASN1 and PEM code to support netscape certificate sequences.
17873 * Add several PKIX and private extended key usage OIDs.
17877 * Modify the 'ca' program to handle the new extension code. Modify
17878 openssl.cnf for new extension format, add comments.
17882 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
17883 and add a sample to openssl.cnf so req -x509 now adds appropriate
17888 * Continued X509 V3 changes. Add to other makefiles, integrate with the
17889 error code, add initial support to X509_print() and x509 application.
17893 * Takes a deep breath and start adding X509 V3 extension support code. Add
17894 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
17895 stuff is currently isolated and isn't even compiled yet.
17899 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
17900 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
17901 Removed the versions check from X509 routines when loading extensions:
17902 this allows certain broken certificates that don't set the version
17903 properly to be processed.
17907 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
17908 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
17909 can still be regenerated with "make depend".
17913 * Spelling mistake in C version of CAST-128.
17915 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
17917 * Changes to the error generation code. The perl script err-code.pl
17918 now reads in the old error codes and retains the old numbers, only
17919 adding new ones if necessary. It also only changes the .err files if new
17920 codes are added. The makefiles have been modified to only insert errors
17921 when needed (to avoid needlessly modifying header files). This is done
17922 by only inserting errors if the .err file is newer than the auto generated
17923 C file. To rebuild all the error codes from scratch (the old behaviour)
17924 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
17925 or delete all the .err files.
17929 * CAST-128 was incorrectly implemented for short keys. The C version has
17930 been fixed, but is untested. The assembler versions are also fixed, but
17931 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
17932 to regenerate it if needed.
17933 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
17934 Hagino <itojun@kame.net>*
17936 * File was opened incorrectly in randfile.c.
17938 *Ulf Möller <ulf@fitug.de>*
17940 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
17941 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
17942 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
17943 al: it's just almost always a UTCTime. Note this patch adds new error
17944 codes so do a "make errors" if there are problems.
17948 * Correct Linux 1 recognition in config.
17950 *Ulf Möller <ulf@fitug.de>*
17952 * Remove pointless MD5 hash when using DSA keys in ca.
17954 *Anonymous <nobody@replay.com>*
17956 * Generate an error if given an empty string as a cert directory. Also
17957 generate an error if handed NULL (previously returned 0 to indicate an
17958 error, but didn't set one).
17960 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
17962 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
17966 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
17967 parameters. This was causing a warning which killed off the Win32 compile.
17971 * Remove C++ style comments from crypto/bn/bn_local.h.
17973 *Neil Costigan <neil.costigan@celocom.com>*
17975 * The function OBJ_txt2nid was broken. It was supposed to return a nid
17976 based on a text string, looking up short and long names and finally
17977 "dot" format. The "dot" format stuff didn't work. Added new function
17978 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
17979 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
17980 OID is not part of the table.
17984 * Add prototypes to X509 lookup/verify methods, fixing a bug in
17985 X509_LOOKUP_by_alias().
17989 * Sort openssl functions by name.
17993 * Get the gendsa program working (hopefully) and add it to app list. Remove
17994 encryption from sample DSA keys (in case anyone is interested the password
17999 * Make *all* `*_free` functions accept a NULL pointer.
18001 *Frans Heymans <fheymans@isaserver.be>*
18003 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18006 *Anonymous <nobody@replay.com>*
18008 * s_server should send the CAfile as acceptable CAs, not its own cert.
18010 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18012 * Don't blow it for numeric -newkey arguments to apps/req.
18014 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18016 * Temp key "for export" tests were wrong in s3_srvr.c.
18018 *Anonymous <nobody@replay.com>*
18020 * Add prototype for temp key callback functions
18021 SSL_CTX_set_tmp_{rsa,dh}_callback().
18025 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18026 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18030 * X509_name_add_entry() freed the wrong thing after an error.
18032 *Arne Ansper <arne@ats.cyber.ee>*
18034 * rsa_eay.c would attempt to free a NULL context.
18036 *Arne Ansper <arne@ats.cyber.ee>*
18038 * BIO_s_socket() had a broken should_retry() on Windoze.
18040 *Arne Ansper <arne@ats.cyber.ee>*
18042 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18044 *Arne Ansper <arne@ats.cyber.ee>*
18046 * Make sure the already existing X509_STORE->depth variable is initialized
18047 in X509_STORE_new(), but document the fact that this variable is still
18048 unused in the certificate verification process.
18050 *Ralf S. Engelschall*
18052 * Fix the various library and apps files to free up pkeys obtained from
18053 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18057 * Fix reference counting in X509_PUBKEY_get(). This makes
18058 demos/maurice/example2.c work, amongst others, probably.
18060 *Steve Henson and Ben Laurie*
18062 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18063 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18064 are no longer created. This way we have a single and consistent command
18065 line interface `openssl <command>`, similar to `cvs <command>`.
18067 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18069 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18070 BIT STRING wrapper always have zero unused bits.
18074 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18078 * Make the top-level INSTALL documentation easier to understand.
18082 * Makefiles updated to exit if an error occurs in a sub-directory
18083 make (including if user presses ^C) [Paul Sutton]
18085 * Make Montgomery context stuff explicit in RSA data structure.
18089 * Fix build order of pem and err to allow for generated pem.h.
18093 * Fix renumbering bug in X509_NAME_delete_entry().
18097 * Enhanced the err-ins.pl script so it makes the error library number
18098 global and can add a library name. This is needed for external ASN1 and
18099 other error libraries.
18103 * Fixed sk_insert which never worked properly.
18107 * Fix ASN1 macros so they can handle indefinite length constructed
18108 EXPLICIT tags. Some non standard certificates use these: they can now
18113 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18114 into a single doc/ssleay.txt bundle. This way the information is still
18115 preserved but no longer messes up this directory. Now it's new room for
18116 the new set of documentation files.
18118 *Ralf S. Engelschall*
18120 * SETs were incorrectly DER encoded. This was a major pain, because they
18121 shared code with SEQUENCEs, which aren't coded the same. This means that
18122 almost everything to do with SETs or SEQUENCEs has either changed name or
18123 number of arguments.
18125 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18127 * Fix test data to work with the above.
18131 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18132 was already fixed by Eric for 0.9.1 it seems.
18134 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18136 * Autodetect FreeBSD3.
18140 * Fix various bugs in Configure. This affects the following platforms:
18144 unixware-2.0-pentium
18149 * Eliminate generated files from CVS. Reorder tests to regenerate files
18150 before they are needed.
18154 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18158 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18160 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18161 changed SSLeay to OpenSSL in version strings.
18163 *Ralf S. Engelschall*
18165 * Some fixups to the top-level documents.
18169 * Fixed the nasty bug where rsaref.h was not found under compile-time
18170 because the symlink to include/ was missing.
18172 *Ralf S. Engelschall*
18174 * Incorporated the popular no-RSA/DSA-only patches
18175 which allow to compile a RSA-free SSLeay.
18177 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18179 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18180 when "ssleay" is still not found.
18182 *Ralf S. Engelschall*
18184 * Added more platforms to Configure: Cray T3E, HPUX 11,
18186 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18188 * Updated the README file.
18190 *Ralf S. Engelschall*
18192 * Added various .cvsignore files in the CVS repository subdirs
18193 to make a "cvs update" really silent.
18195 *Ralf S. Engelschall*
18197 * Recompiled the error-definition header files and added
18198 missing symbols to the Win32 linker tables.
18200 *Ralf S. Engelschall*
18202 * Cleaned up the top-level documents;
18203 o new files: CHANGES and LICENSE
18204 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18205 o merged COPYRIGHT into LICENSE
18206 o removed obsolete TODO file
18207 o renamed MICROSOFT to INSTALL.W32
18209 *Ralf S. Engelschall*
18211 * Removed dummy files from the 0.9.1b source tree:
18212 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18213 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18214 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18215 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18216 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18218 *Ralf S. Engelschall*
18220 * Added various platform portability fixes.
18224 * The Genesis of the OpenSSL rpject:
18225 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18226 Young and Tim J. Hudson created while they were working for C2Net until
18229 *The OpenSSL Project*
18231 ### Changes between 0.9.0b and 0.9.1b [not released]
18233 * Updated a few CA certificates under certs/
18237 * Changed some BIGNUM api stuff.
18241 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18242 DGUX x86, Linux Alpha, etc.
18246 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18247 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18252 * Add -strparse option to asn1pars program which parses nested
18255 *Dr Stephen Henson <shenson@bigfoot.com>*
18257 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18261 * DSA fix for "ca" program.
18265 * Added "-genkey" option to "dsaparam" program.
18269 * Added RIPE MD160 (rmd160) message digest.
18273 * Added -a (all) option to "ssleay version" command.
18277 * Added PLATFORM define which is the id given to Configure.
18281 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18285 * Extended the ASN.1 parser routines.
18289 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18293 * Added a BN_CTX to the BN library.
18297 * Fixed the weak key values in DES library
18301 * Changed API in EVP library for cipher aliases.
18305 * Added support for RC2/64bit cipher.
18309 * Converted the lhash library to the crypto/mem.c functions.
18313 * Added more recognized ASN.1 object ids.
18317 * Added more RSA padding checks for SSL/TLS.
18321 * Added BIO proxy/filter functionality.
18325 * Added extra_certs to SSL_CTX which can be used
18326 send extra CA certificates to the client in the CA cert chain sending
18327 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18331 * Now Fortezza is denied in the authentication phase because
18332 this is key exchange mechanism is not supported by SSLeay at all.
18336 * Additional PKCS1 checks.
18340 * Support the string "TLSv1" for all TLS v1 ciphers.
18344 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18345 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18349 * Fixed a few memory leaks.
18353 * Fixed various code and comment typos.
18357 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18358 bytes sent in the client random.
18360 *Edward Bishop <ebishop@spyglass.com>*
18364 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18365 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18366 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18367 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18368 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18369 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18370 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18371 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18372 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18373 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18374 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18375 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18376 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18377 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18378 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18379 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18380 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18381 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18382 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18383 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18384 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18385 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18386 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18387 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18388 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18389 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18390 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18391 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18392 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18393 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18394 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18395 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18396 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18397 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18398 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18399 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18400 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18401 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18402 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18403 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18404 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18405 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18406 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18407 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18408 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18409 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18410 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18411 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18412 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18413 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18414 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18415 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18416 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18417 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18418 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18419 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18420 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18421 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18422 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18423 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18424 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18425 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18426 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18427 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18428 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18429 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18430 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18431 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18432 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18433 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18434 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18435 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18436 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18437 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18438 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18439 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18440 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18441 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18442 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18443 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18444 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18445 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18446 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18447 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18448 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18449 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18450 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18451 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18452 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18453 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18454 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18455 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18456 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18457 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18458 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18459 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18460 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18461 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18462 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18463 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18464 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18465 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18466 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18467 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18468 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18469 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18470 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18471 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18472 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18473 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18474 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18475 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18476 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18477 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18478 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18479 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18480 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18481 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18482 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18483 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18484 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18485 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18486 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18487 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18488 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18489 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18490 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18491 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18492 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18493 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18494 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18495 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18496 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18497 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18498 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18499 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18500 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18501 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18502 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18503 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18504 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18505 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18506 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18507 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18508 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18509 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18510 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18511 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18512 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18513 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18514 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18515 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18516 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18517 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18518 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18519 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18520 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18521 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18522 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18523 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655